Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The C.P.R.’s capacity and investment strategy in Rogers Pass, B.C., 1882-1916 Backler, Gary G. 1981

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
UBC_1981_A4_6 B33.pdf [ 21.01MB ]
Metadata
JSON: 1.0095109.json
JSON-LD: 1.0095109+ld.json
RDF/XML (Pretty): 1.0095109.xml
RDF/JSON: 1.0095109+rdf.json
Turtle: 1.0095109+rdf-turtle.txt
N-Triples: 1.0095109+rdf-ntriples.txt
Original Record: 1.0095109 +original-record.json
Full Text
1.0095109.txt
Citation
1.0095109.ris

Full Text

THE C.P.R.'S CAPACITY AND INVESTMENT STRATEGY IN ROGERS PASS, B.C., 1882-1916. y^-yGary G. Backler, B. A., Oxon., 1976. A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE (BUSINESS ADMINISTRATION) THE FACULTY OF GRADUATE STUDIES Faculty of Commerce and Business Administration We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA By in May 1981 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of cv^r-ieg-c^e; AroD SuSi^ES^ AOHi^iSTicpnorvi The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date np-fi 17/19) i i ABSTRACT CP Rail is currently confronted by a capacity problem on its main line in Rogers Pass, at the summit of the Selkirk Mountains. The single-track, steeply graded facility is inadequate for the forecasted demand for traffic flows in the westbound direction. The Company must decide whether to continue to operate over the present line, incurring high operating costs and escalating congestion costs, or whether to invest in an 8.9-mile tunnel which, by reducing the gradient against westbound traffic, will stem congestion and reduce the level of operating costs in future years. CP Rail must make a trade-off between construction costs and operating costs. The Company has made such a trade-off in Rogers Pass on at least two previous occasions. The first occasion was that prior to completion of the initial transcontinental rail link, when the decision was taken to breach the Selkirk Mountains by a surface crossing through Rogers Pass. The second occasion was that prior to the decision, taken in 1913, to abandon the surface alignment through Rogers Pass in favour of a five-mile tunnel beneath the summit of the Selkirks. This thesis identifies the factors which impelled the taking of trade-off decisions in each situation, allocates an appropriate weighting to the factors, and examines the criteria upon which the investment decisions were based. Previous historians of the C.P.R.'s operations in Rogers Pass emphasise the influence of avalanches upon the investment decisions taken. Many of these historians interpret the C.P.R.'s surface operations as an unremitting campaign to protect its traffic against snowslides, and they interpret the Company's decision to construct the Connaught Tunnel as an acknowledgement of defeat in the campaign. This thesis emphasises th°e economic and commercial aspects of the C.P.R.'s operations in Rogers Pass, and a quantitative approach towards the analysis is adopted. Part One of the thesis is concerned with the initial decision to construct the C.P.R. main line across the surface of the Selkirks through Rogers Pass. Part Two is concerned with the decision to abandon the surface alignment. Part One begins with an explanation of the engineering and economic problems of locating railway lines through mountainous terrain, and examines how these problems were handled by the C.P.R. in the specific circumstances of Rogers Pass. The expectations of the railway builders for construction and operation through the Pass are compared with the realities which were encountered. Analysis reveals that the gap between expectations and realities was not wide, and that the surface alignment adopted by the C.P.R. provided an adequate, economical solution to the problem of breaching the Selkirk Mountains by rail, at least until the turn of the 20th Century. Part Two begins with an analysis of the influence of avalanches in Rogers Pass upon the decision to relocate the main line underground. The analysis strongly suggests that neither the 1910 avalanche disaster in particular, nor the cost of protecting traffic from snowslides in general, were sufficient to justify investment in the Connaught Tunnel. An examination of the operating conditions, traffic growth iv and traffic forecasts through the Selkirk Mountains in the early years of the 20th Century reveals that the C.P.R. faced high operating costs and escalating congestion costs on the surface route by 1913. The Company had already invested in system improvements elsewhere in the mountains in order to reduce these costs. Confronted by the inadequacy of its existing facility for forecasted demand in Rogers Pass, the C.P.R. decided in 1913 to drive a double-track tunnel beneath the summit of the Selkirks, and to abandon the surface route. Analysis of the C.P.R.'s evaluations of alternative proposed tunnels confirms that the principal economic benefit of the project was the savings in train-haulage costs, and not the savings in the cost of avalanche defence. CONTENTS ABSTRACT i i LIST OF TABLES viii LIST OF ILLUSTRATIONS x ACKNOWLEDGEMENTS xChapter 1 INTRODUCTION 1 Objectives 4 Scope of the Thesis 6 Outline . .' 7 Data Sources and Limitations 9 PART ONE UP AND OVER 13 2 RAILWAYS AND MOUNTAINS 15 2.1 The Two Solutions(a) "Low Capital Cost, High Operating Cost" Solution 17 (i) Gradients 20 (ii) Curvature 4 (b) "High Capital Cost, Low Operating Cost" Solution 27 2.2 The Trade-Off 28 3 RAILWAYS AND ROGERS PASS 3 5 3.1 Rogers Pass 35 (a) Location 7 (b) Topography(c) Climate 9 3.2 The Selection of Rogers Pass for the First Transcontinental Rail Link 42 3.3 The Expectations of the Builders 50 (a) The Character of Construction Work ... 51 (b) The Cost of Construction 52 (c) Time Required for Construction 53 (d) Operating Methods and Traffic Forecasts 54 (e) Snowslide Protection 56 4 REALITIES 68 (a) The Character of Construction Work 69 (b) The Cost of Construction 75 (c) Time Required for Construction 78 (d) Operating Methods and Traffic Flows 81 (e) Snowslide Protection 8vi PART TWO THE BIG BORE Ill . 5 AVALANCHE PROBLEMS 114 5.1 The 1910 Disaster 115 5.2 The Snow Problem In General 123 (a) The Direct Costs of Maintaining the Avalanche Defence System 123 (b) The Indirect Cost of Disruptions to Traffic 129 (i) The Nature of Disruption Costs 130 (ii) The Incidence of Disruption .. 133 (iii) Diversionary Arrangements .... 156 (iv) Was Disruption Increasing? ... 158 (v) Disruption Costs and the Abandonment Decision 161 6 CAPACITY PROBLEMS 173 6.1 The Capacity of the Main Line 173 (a) Train Weight 174 (b) Train Paths 182 6.2 Traffic Flows 185 (a) Total Traffic Levels 18(b) Changes in Specific Traffic Flows .... 191 (i) Passenger 19(ii) Lumber 5 (iii) Grain 198 (iv) Fish 202 (v) Other Transcontinental Traffic 205 (vi) Local Traffic 206 6.3 Competitive Pressures 207 (a) C.P.R. Rates in the Mountains 20(b) The Sources of Competitive Pressure .. 210 (c) The C.P.R.'s Perception of the Pressures 214 6.4 "System" Improvements to the C.P.R 220 (a) Large-Scale Improvements Beyond the Selkirks 221 (i) The Ottertail Diversion 22(ii) The Palliser Tunnel 222 (iii) The Spiral Tunnels 3 (iv) C.P.R. Investment Strategy in the Rockies 230 (b) Smaller-Scale Improvements Within the Selkirks 235 (i) Improvements to Rolling Stock. 235 (ii) Improvements to Infrastructure 242 (iii) C.P.R. Investment Strategy in the Selkirks 249 vii 6.5 The Financial Resources of the C.P.R 258 6.6 Traffic Forecasts and their Implications ... 259 7 ALTERNATIVES AND THEIR EVALUATION 285 7.1 Alternatives Beyond the Selkirk Mountains .. 285 (a) Alternatives south of Rogers Pass .... 286 (b) The Yellowhead Pass 288 7.2 Alternatives Within the Selkirk Mountains .. 289 7.3 Alternative Tunnels 301 (a) The Kilpatrick Tunnel 302 (b) The Busteed Tunnel 317 (c) The Sullivan Tunnel 324 7.4 Criteria and Objectives 5 8 THE CONNAUGHT TUNNEL 338 8.1 The Alignment as Contracted 339 8.2 A "Social Cost-Benefit Analysis" of the Contracted Alignment 353 8.3 The Alignment as Constructed 364 9 CONCLUSIONS 39Suggestions for Further Research 407 SELECTED BIBLIOGRAPHY 414 I vi i i LIST OF TABLES 1. Siding Accommodation in the Selkirk Mountains, c. 1896 87 2. Passenger Train Record, Mountain Subdivision, 1908. 137 3. Average Number of Trains Per Day, Mountain and Shuswap Sections, 1906-1908 140 4. Average Weight of Trains, Mountain and Shuswap Sections, 1906-1908 143 5. Total Equivalent Gross Tonnage Per Month, Mountain and Shuswap Sections, 1906-1908 146 6. Comparison of Traffic on Mountain, and Shuswap Sections, Slide Seasons (January-April), 1906-1908 149 7. Total Equivalent Gross Ton Mileage Per Month, Mountain Subdivision, 1910 and 1911 152 8. Comparison of Equivalent Gross Ton Mileages, Mountain Subdivision, 1910 and 1911 153 9. Tonnage Ratings for single 210% locomotive between stations on Mountain Subdivision, prior to June 1913 175 10. Average Train Weights, Mountain Subdivision, 1906-1913 9 11. Gross Tonnage of Passenger and Freight Traffic over each mile of road, Mountain and Shuswap Subdivisions, 1904-1913 186 12. Annual Rates of Change in Gross Tonnage per mile, Mountain and Shuswap Subdivisions, 1904-1913 187 13. Balance of Traffic Flows through Rogers Pass, 1889-1913 9 14. Passenger Volume through Rogers Pass, various months, 1893-1908 192 15. Lumber Traffic through Rogers Pass, various years, 1900-1918 196 16. Grain Traffic through Rogers Pass, various years, 1903-1917 200 17. B.C. Salmon Production and Tra-de, 1887-1918 203 18. The Allocation of Infrastructure Investment on the C.P.R., 1901-1913 232 19. Bridge Improvements in the Selkirk Mountains, 1893-1909 245 20. Proportion of Total C.P.R. Freight Traffic handled over Selkirk Mountains, 1904-1913 252 21. Annual Rates of.Change in Regional Distribution of Freight Traffic, 1904-1913 253 22. Cost Comparison of Double-Tracking Alternatives through the Selkirk Mountains, October 1912 .... 294 ix 23. Results of Cost Analyses of Alternative Investments in Rogers Pass, 1912-13 299 24. Cost-Benefit Analysis of Kilpatrick's Proposed Tunnel Alignment of May 1912 312 25. Cost-Benefit Analysis of Busteed's Proposed Tunnel Alignment of October 1912 320 26. Comparison of Percentage Tenders of Cost Per Foot of Rock Section, Rogers Pass Tunnel 346 27. Cost-Benefit Analysis of Connaught Tunnel Alignment, as Contracted for, June 1913 348 28. Revised Estimate of Benefits of Contracted Tunnel Alignment 363 29. Grain Exports from Vancouver, 1910-1935 378 30. Costs and Benefits of the Cancellation of the East-Slope Revision 381 31. Cost-Benefit Analysis of Connaught Tunnel Alignment, as Completed, December 1916 383 x LIST OF ILLUSTRATIONS Map I. C.P.R. Main Line from Revelstoke to Laggan II. Location of Alternative Alignments Proposed at Construction Time on the C.P.R. Main Line in Rogers Pass III. Location of Alternative Alignments and Tunnels on the C.P.R. Main Line in Rogers Pass Figure 1. Profile of C.P.R. Main Line between Revelstoke and Beavermouth, showing Tonnage Ratings for single 210% locomotive before and after Dynamometer Tests, May 1913 2. Profiles of Alternative Alignments and Tunnels on the C.P.R. Main Line in Rogers Pass 36 71 304 180 305 l xi ACKNOWLEDGEMENTS I would like to express at the outset my deepest gratitude to Dr. T. D. Heaver, Chairman of the Transportation Division at the University of British Columbia, for his motivation of this project, and for his sustained interest in its progress. I would also like to thank the members of my thesis committee for their ready guidance of my research efforts. For their assistance in making available indispensable archival material, I wish to extend thanks to each and all of the following:- 0. S. A. Lavallee, James Shields, and the staff of the Canadian Pacific Corporate Archives in Montreal; Dr. Carl Vincent, and the staff of the Public Archives of Canada; John Woods, and the staff of Mount Revelstoke and Glacier National Parks; Dave Lightheart, and CP Rail, Vancouver; the staff of the Glenbow Alberta Institute; the staff of the Public Archives of British Columbia and the Legislative Library; the staff of the Vancouver City Archives; and the staff of the Special Collections division at U.B.C. In addition, I would like to offer especial thanks to Mr. and Mrs. Donald Kilpatrick, here in Vancouver, for their generosity in providing me with access to the private papers of Thomas Kilpatrick, and for their generosity in providing me, on more than one occasion, with an excellent supper. xii To ANDREA, "0 how I long to travell back And tread again that ancient track..." Henry Vaughan, "The Retreate" 1 CHAPTER 1 INTRODUCTION Rail access to the west coast of Canada is rendered expensive by the topography of British Columbia, where four major mountain ranges intrude between the western seaboard and the eastern boundary of the province. In constructing railways through B.C., the builders have always had to face the dilemma of either investing large sums of capital per mile of line in order to obtain an easy alignment over which traffic may flow smoothly, or building to inferior standards and subsequently incurring high operating costs in the movement of trains. Clearly, these trade-offs between construction costs and operating costs, and between immediate costs and delayed costs, must be made in all railway construction, and indeed in the provision of any transportation infrastructure. Nevertheless, the trade-offs are particularly crucial, and the dilemma is particularly acute, in regions of rugged terrain, where constructional and operating characteristics, and therefore costs, often differ markedly between alternative routes. CP Rail is presently confronted by such a dilemma as it seeks a solution to its next major main-line capacity problem. This problem is posed by the 8.16-mile ascent from Rogers to Stoney Creek, on the eastern approach to Rogers Pass, B.C., at the summit of the Selkirk Mountains. The ascent is over single track, and at a maximum gradient of 2.2 per cent, against westbound traffic for most of the distance. The elevation gained 2 in the 8.16 miles is 899.9 feet. The heaviest westbound trains, already powered by eight locomotives from Golden, require the assistance of an additional five pusher locomotives between Rogers and Stoney Creek. Each train must be brought to a complete stand when the pushers are inserted, and again when the pushers are switched out after the ascent. This necessity to stop the trains twice within ten miles, together with the necessity of returning the pusher locomotives light down the single track against the prevailing flow of traffic, restricts the capacity of the main line. Moreover, the single-track configuration of the existing main line between Beavermouth, east of the Selkirk summit, and Glacier, to the west, contributes to delays in the meeting and passing of trains, further restricting main-line capacity. CP Rail is therefore contemplating the construction of a second main track across the summit of the Selkirks, between Rogers and a point 3.1 miles west of Glacier. The maximum gradient of this track would be one per cent, against westbound traffic, and the maximum angle of curvature would be six degrees. Such an alignment would dispense with the necessity for pusher locomotives, and would resolve the conflict between eastbound and westbound flows, since eastbound traffic would continue to travel over the present line. However, in order to maintain the stipulated gradient and curvature, construction of this alternative route would require the driving of an 8.9-mile tunnel beneath Rogers Pass. The cost of the realignment project is estimated at $300 million in 1980 dollars.1 Moreover, the financial viability of the project is rendered questionable by 3 the uncertainty of future traffic volumes and by the unremunerative nature of the grain traffic, which constitutes some fifteen per cent, of CP Rail's westbound flow by weight. CP Rail, then, is confronted by a dilemma at the summit of the Selkirk Mountains. It must decide whether to continue to operate over the present line, incurring high operating costs and escalating congestion costs, or whether to undertake a massive, indivisible capital investment intended in future years to stem congestion and to reduce permanently the level of operating costs. This decision situation has two distinct facets, and these may be framed in interrogative form. First, what is the alignment which is appropriate to the traffic flows through the mountains? Then, second, what is the appropriate level of investment which should be undertaken in order to secure this alignment? Only after these two fundamental questions have been answered can a decision be reached concerning the adoption of the project. This is not the first occasion upon which Canadian Pacific Railway management have sought answers to these questions. Indeed, they have been forced to address the questions on at least two previous occasions. The first occasion was that prior to completion of the initial transcontinental rail link, when the decision was taken to breach the Selkirk Mountains by a surface crossing over the summit through Rogers Pass. The second occasion, some thirty years later, in 1913, was that prior to the decision to abandon the surface alignment through Rogers Pass in favour of a five-mile tunnel beneath the Pass. This was the Connaught Tunnel, and it is by means of this 4 tunnel that all of CP's main-line traffic still crosses the Selkirk Divide. Objectives Prompted by the resurgence of interest in these questions today, the first objective of this thesis is to undertake a study of these previous comparable decision situations. Such a study will identify the factors which impelled the taking of decisions in each situation, the forces and influences which made the taking of decisions necessary. From the study of these historical situations, the reader will be able to draw his own comparisons with the present decision situation, and will be free to infer his own "lessons to be learned" for the current realignment proposal. The second objective of this thesis, springing directly from the first, is to allocate an appropriate weighting to the factors which influenced the decisions in each of the historical situations. Historians of CP Rail's operations in Rogers Pass have hitherto always emphasised the influence of avalanches upon the decisions taken, and especially upon the decision taken in 1913 to abandon the Pass in favour of an underground route.2 The economics of CP Rail's operations in the Pass, and specifically the economics of the decision to construct the Connaught Tunnel, have been consistently neglected. This neglect is especially serious because reference to the primary documentation which surrounded the decision suggests that the occurrence of avalanches over the surface alignment in Rogers Pass is not alone sufficient explanation for the decision to relocate the main line underground. The economics of the rail operation, and in particular the emergence of a major capacity problem in the Pass, appear then, as now, to have been factors of considerable significance in shaping the decision. This thesis will therefore seek to remedy the neglect of those factors. Fresh evidence will be analysed, and appropriate conclusions drawn. The results will be of interest to anyone who has a concern for historical accuracy. The third and final objective of this thesis is to examine critically the contemporary techniques of appraisal which the Canadian Pacific Railway Company employed in their investment decisions in Rogers Pass during the first thirty-five years of the route's history, and the criteria upon which those decisions were based. The emphasis in this examination will be placed primarily upon the decision to invest in construction of the Connaught Tunnel. Throughout the analysis, however, stress will be placed upon the central and recurring importance of the trade-offs between construction costs and operating costs and between immediate costs and delayed costs, as they applied to rail operations in Rogers Pass. Attention will also be drawn to the manner in which these trade-offs were handled during the course of the study period. The results of this examination of the C.P.R.'s approach to previous investment decisions in Rogers Pass will be of interest to the business historian, and of relevance in a consideration of the evaluation procedures employed by CP Rail today. 6 Scope Of The Thesis Previous studies of the history of the C.P.R. in the Selkirk Mountains may be grouped into two general categories, biographical and descriptive. The biographical studies have concentrated upon the personalities involved in the decisions to locate, construct and operate the railway through Rogers Pass.3 The descriptive studies have tended either to concentrate upon the construction phase of the line,4 or to treat the line, once built, as an incidental factor in the discussion of regional development.5 The present study is essentially analytical in character. It is an historical study of railway economics and their investment implications for a particular operating situation, that of the C.P.R. in Rogers Pass. Therefore, it differs from previous studies in both emphasis and focus. The emphasis in this study is upon the economic and commercial aspects of railway operation in Rogers Pass, and a quantitative approach is therefore adopted. Such an approach permits integration into the analysis of quantitative data, data which has hitherto received little consideration from historians and analysts, and yet which, as has been indicated above, suggests an alternative interpretation of the C.P.R.'s investment decisions in Rogers Pass to that which is common currency. The focus of this study is intentionally narrow and highly localised. This is a detailed study of railway economics on a forty-five mile stretch of the Canadian Pacific main line, a line nearly three thousand miles long. The study opens in 1882, I 7 the year in which Rogers Pass was selected as part of the route for the transcontinental railway, and it closes in 1916, the year in which the Connaught Tunnel was opened to traffic, marking the abandonment of surface rail operations through the Pass. Throughout that period, railway investment decisions and the economics of railway operations will be analysed and evaluated. Wider historical developments contemporaneous with the study period, such as the economic advance of British Columbia, the opening up of the prairies, and the construction of additional transcontinental routes, by both rail and water, will be considered only insofar as they impinged upon the economics of C.P.R.'s operations in Rogers Pass. Outline The thesis is divided into two parts. The first part (Chapters 2-4) is primarily concerned with the inception and realisation of a rail link across the Selkirk Mountains, accomplished by means of a surface alignment over Rogers Pass. This section embraces the construction of the original line and the measures taken to consolidate the link as a feasible route for trans-mountain traffic. The second part (Chapters 5-9) analyses the forces of change, economic and non-economic, which led to questioning of the appropriateness of the surface alignment, and ultimately to the identification of a decision problem. The analysis proceeds to examine the generation and evaluation of alternative solutions to the problem, and concludes with an appraisal of the achievement of the alternative which was implemented, that of relocating the 8 line underground through the Connaught Tunnel. Part One begins with a "layman's guide" to the engineering problems of locating railway lines through mountainous terrain, and to the implications of those problems for the economics of railway operations. Pertinent engineering concepts are elucidated. The nature of the critical trade-off between construction costs and operating costs is explained. Chapter 3 describes the physical and climatic characteristics of the Selkirk Mountains, and the implications of these characteristics for railway location and operation. The reasons for the selection of Rogers Pass as the route by which the C.P.R. would cross the Selkirk Mountains are explained. The expectations of the railway builders for both construction and operation through the Pass, expectations which led to the adoption of the route, are made explicit. In Chapter 4, the realities of construction and operation over the surface alignment are described, and the gaps between expectations and realities are highlighted. Measures intended to close these gaps are reviewed, and their success evaluated. Part Two commences, in Chapter 5, with an analysis of the influence of avalanches in Rogers Pass upon the decision to relocate the main line underground. The widely accepted interpretations of the 1910 avalanche disaster in particular and of the snowslide problem in general are presented. These interpretations are then challenged, and as a result of this challenge an alternative interpretation is offered of the influence of avalanches upon the decision to abandon the surface alignment. 9 Chapter 6 contains an analysis of traffic developments through the mountains of B.C. The analysis includes consideration of changes in traffic volume and composition, competitive pressures, and the cumulative results of improvements undertaken elsewhere on the C.P.R. system. The analysis suggests the increasing inadequacy of the surface alignment over Rogers Pass to cope with these developments: a decision problem is identified and specified. Chapter 7 reviews the alternative solutions which were generated for this decision problem, and presents the results of the screening of these alternatives. In Chapter 8, a detailed evaluation of the preferred alternative is carried out. This alternative was a realignment which included the boring of the Connaught Tunnel. The project was modified during implementation: the scheme as conceived and the scheme as completed are both appraised. In the final chapter, the conclusions of the thesis are presented, and suggestions are offered for further research. Data Sources And Limitations This thesis is based upon three primary sources of unpublished data. The first is the Letterbooks and inward correspondence of the C.P.R. Presidents. The Letterbooks have been microfilmed, but the inward correspondence is available only at the Canadian Pacific Corporate Archives in Windsor Station, Montreal. Although these Archives were only recently established, and although the task of indexing the vaults of corporate records is not far advanced, two complete files on 10 "The Rogers Pass Tunnel" have been assembled. Whilst weak in quantitative content, the files provide invaluable insight into the objectives of the project and the stages of its implementat ion. The second source is the collection of C.P.R. records and correspondence which is held at the Revelstoke City Museum, B.C. This miscellaneous collection, withheld from the Canadian Pacific Corporate Archives, was comprehensively indexed in 1976. It includes complete monthly records of traffic volumes on the First District of the British Columbia Subdivision, of which Rogers Pass was a part, for the years 1910 and 1911, and sporadic records of expenditures on snow sheds and snow clearance for certain months between 1912 and 1914. The third source is the personal notebooks and diaries of Thomas Kilpatrick, "The Snow King," who ended a thirty-year career with the C.P.R. in the mountains of B.C. as Superintendent of the Mountain and Shuswap Sections from 1901 to 1912. His papers are now held by his son, Mr. Donald Kilpatrick, as part of a private collection preserved in Vancouver, B.C. The documents contain many fascinating details of railway maintenance and operations in the mountains throughout the study period. Of particular value in the preparation of this thesis was a monthly record, maintained uninterrupted throughout the years 1906 to 1908, of the number of main-line trains per day, their weights and transit times. In comparison with the inter-war period, the years prior to the First World War are rich in documentary sources concerning the management of the C.P.R. Nevertheless, it is a sad fact 11 that much information has been destroyed. The loss is particularly serious for a sharply focussed project such as the present thesis which attempts to apply quantitative techniques to the analysis of an investment decision which was undertaken some seventy years ago. Detailed information on the costs of operating trains over Rogers Pass may never have existed. If it did, it has certainly not survived, and neither has detail of the anticipated costs and benefits of the several realignment schemes which were proposed in 1912. Similarly, very little data exists concerning traffic flows through the mountains either by volume, commodity or direction of movement. Assembly of the extant data was a piecemeal and painful process, drawing from many diverse sources. Many gaps remain. Where possible, the gaps have been filled by inference and assumption; where not, they have simply been identified and recorded. It is to be hoped that this thesis will at least prevent the gaps from widening any further. 12 FOOTNOTES 1 "Midweek Report," Vancouver Province, April 16, 1980, p. D 1. 2 See chapter 5. 3 See, for example, W. Vaughan, The Life And Work Of Sir William  Van Home, New York: The Century Co., 1920; H. Gilbert, Awakening Continent, Aberdeen: Aberdeen University Press, 1965, and, The End Of The Road, Aberdeen: Aberdeen University Press, 1977; C. A. Shaw, Tales Of A Pioneer Surveyor, Toronto: Longman, 1970. 4 See, for example, 0. S. A. Lavallee, Van Home' s Road, Montreal: Railfare Enterprises, 1975; P. Berton, The Last Spike:  The Great Railway 1881-1885, Toronto: McClelland and Stewart, 1971. 5 See for example, J. S. Marsh, "Man, Landscape And Recreation In Glacier National Park, British Columbia, 1880 to present," PhD thesis, University of Calgary, 1971; W. W. Bilsland, "A History Of Revelstoke And The Big Bend, " MA thesis, University of British Columbia, 1955. 13 PART ONE UP AND OVER This part of the thesis is concerned exclusively with the surface alignment of the C.P.R. main line through Rogers Pass. A brief introduction is provided to the engineering and commercial problems of locating and operating railways in mountains. Then, the specific engineering and commercial problems of locating and operating the C.P.R. main line over the summit of the Selkirk Mountains are analysed. The analysis compares the expectations and the realities of the C.P.R. for specific areas of concern along the surface alignment, and evaluates measures undertaken in order to promote correspondence between those expectations and realities. Part One concludes with a consideration of the extent to which such correspondence was achieved. The analysis in Part One spans the selection, construction and operation of the Rogers Pass route until the turn of the century, that is, prior to the occurrence of the 1910 avalanche disaster. Consideration of this disaster initiates the analysis in Part Two. A division of the thesis prior to the 1910 disaster is maintained in deference to previous historians of the C.P.R.'s operations in Rogers Pass. Most of these historians identify the 1910 disaster as a turning point in C.P.R. management's perception of the viability of the surface alignment. The validity of this identification is challenged directly in Part Two. However, the foundation for this challenge is provided by the results of the analysis in Part One. There are three advantages of structuring the analysis in 14 this way. First, it permits establishment in Part One of the extent of the gaps between the expectations and realities of the C.P.R.'s concerns with the surface alignment, before proceeding in Part Two to consider changes in operating conditions which may have widened these gaps. Second, it permits both the avalanche problems and the other operating problems of the line to be considered each as a continuum, stretching from the opening of the surface alignment to the opening of the Connaught Tunnel. Both the avalanche and the operating problems are shown to have changed in severity but not in nature throughout the years of the summit route. Third, it permits the remedial measures undertaken in response to the avalanche problems and the other operating problems to be considered as continua also. Thus, it highlights the extent of substantive changes in the responses of the C.P.R. to the changes in operating conditions over time. It is therefore possible, as a result of adopting this structure, to determine whether or not the 1910 avalanche disaster was indeed a turning point, and whether or not it is at all even meaningful to talk of "turning points" in the context of the C.P.R.'s investments in improving the surface alignment over Rogers Pass. 15 CHAPTER 2 RAILWAYS AND MOUNTAINS This chapter provides a theoretical underpinning to the character of the investment decisions which are analysed in the remainder of the thesis. The chapter begins with an explanation of the two generic types of investment solution to the problem of penetrating mountainous terrain with railway tracks. The fundamental engineering principles involved in each type of solution are described, and the implications of these principles for the economics of railway construction and operation are considered. The inverse relationship between construction costs and operating costs is explained, and criteria are suggested for the trade-off decision between the two types of costs.1 2.1 The Two Solutions The physical barrier which mountains pose to all forms of human communication translates into an economic barrier, as the cost of providing and maintaining communication across mountainous terrain. For rail, the essence of the physical barrier is that the low level of friction between wheel and rail, which affords economic advantage in traffic movement over an even alignment, manifests itself as a low level of adhesion on an adverse gradient, and is therefore turned to economic disadvantage since the payload which can be hauled by a single locomotive is reduced in proportion to the increase in the adversity of the gradient. The following table, adapted from A. M. Wellington's authoritative "The Economic Theory Of The 16 Location Of Railways," demonstrates the reduction in payload which is caused by increasingly adverse gradients. The payloads are those which were capable of being hauled by a single "Standard Heavy Consolidation" steam locomotive, as recorded in 19152: Gradient Payload (per cent. ) (tons) Level • 2,920 0.2 1,920 0.4 1,420 0.6 1,120 0.8 921.0 777 1.2 670 1.4 587 1.6 520 1.8 465 2.0 420 2.2 382 4.5 165 Conceptually, there are two ways in which the physical barrier of mountains may be overcome by railway operation. On the one hand, the ratio of motive power to payload may be adjusted, in order to ensure that traffic can be hauled over the existing adverse gradient. Such an adjustment is effected, either by increasing the number and power of the locomotives 17 which haul the train, or by reducing the weight of the train, or by a combination of these approaches. On the other hand, the existing adverse gradient may be eliminated, in order to ensure that no reduction in the level of adhesion between wheel and rail occurs, and therefore, that no adjustment in the ratio between motive power and payload is necessary. An elimination of adverse gradient is effected, either by reducing the absolute height over which the traffic must be lifted (through tunnelling, cutting, or curvature around the summit) or by reducing the rate at which that height is attained (through "development" of the line, that is, the insertion of length, and usually of curvature, over the summit) or, again, by a combination of these approaches. In order to explore the implications of these engineering solutions for the economics of the railway operation, it is appropriate to classify the solutions according to economic concepts. The first solution, of adjusting the ratio between motive power and payload, may be characterised as a "Low capital cost, high operating cost" solution, and the second solution, that of eliminating the adverse gradient, may be characterised as a "High capital cost, low operating cost" solution. 2.1 (a) "Low Capital Cost, High Operating Cost" Solution. The capital requirement for construction of a line of railway may be decreased by locating the tracks over the natural alignment of the terrain in a route which minimises the necessity for man-made structures. In mountainous regions, location in accordance with this principle may involve either 18 steep gradients or protracted curvature or both. Nevertheless, once the location has been decided, the construction period, over which capital disbursements are spread, will be short relative to the period over which interest on the capital will be repaid, or relative to the period over which railway operations will be carried out upon the line. The capital cost of constructing the line may therefore be regarded as an "immediate" cost: it is the cost which is "immediately" incurred in providing the rail facility. When construction of the facility has been completed, and railway operations have commenced, every train which traverses the route is required to negotiate the steep gradients and protracted curvature, and thereby incurs higher operating costs than would have been incurred with lesser gradients and curvature. Hence, the operating costs of the facility may be regarded as being inversely related to its construction costs. The higher operating costs will be spread over the entire period during which rail operations are carried out over the low-capital-cost alignment. These operating costs may be regarded as "delayed," not only until construction is completed, but until it is necessary to run a train over the line. In the extreme case, if no trains are run over the line, the higher operating costs are "delayed" for perpetuity. It is important to note that although the operating costs of the facility are inversely related to its construction costs, the level of operating costs which is associated with a particular level of investment in construction is not necessarily uniform. The level of operating costs is also 19 related by a more complex function to the volume of traffic which uses the facility. It is helpful to portray the costs of the facility as a U-shaped curve, the locus of which is determined by variations in traffic volume. For even the most cheaply built railway, the level of operating costs may at first decline with increasing traffic, as surplus line capacity is absorbed, equipment better utilised, and the costs of providing motive power and manpower and of maintaining lineside structures are spread over a greater volume of business. However, as traffic continues to increase, and as the surplus capacity provided by the opening of the facility is absorbed, operating costs may begin to rise with the expansion of business, and the marginal rate of increase of the operating costs may exceed the marginal rate of increase of the traffic volume. This behaviour of the operating-cost curve may result from three factors, which may obtrude singly or simultaneously. First, existing resources of locomotives and traincrews may not be sufficient to handle an increase in traffic. Therefore, the marginal increase in traffic entails increases in motive power and manpower which may be particularly severe on a low-capital-cost, steeply graded alignment where the ratio of motive power to payload is initially low.3 Second, the increase in traffic may cause congestion of the facility, particularly in the case of a low-capital-cost, single-track railway where through speeds are low and where limited passing accommodation is provided. This congestion will increase fuel consumption and power requirements, and limit the absolute volume of traffic handled by the facility. As congestion increases further, the 20 operating-cost curve may bend back upon itself, with the cost of operations continuing to escalate while the volume of business conducted declines below that which it would have been on an uncongested facility. Third, the cost of maintaining lineside structures such as trestle bridges and snowsheds may increase as the structures are subjected to greater weights and frequencies of trains. There may be many such structures on a low-capital-cost, rapidly constructed alignment, and the wear upon them due to pounding from locomotives may be particularly great where the ratio of motive power to payload is low.4 Regardless of these fluctuations in the volume of traffic, the relative level of operating costs which is associated with a low-capital-cost alignment is generally higher than the level of operating costs associated with a capital-intensive alignment. There are two reasons for this, which are related to the gradients and the curvature of the alignment, i) Gradients In order to appreciate the impact of gradients upon the costs of operation over a particular facility, the concept of "gradient systems" must first be understood. Trains are assembled, and motive power allocated, according to the length and steepness of particular gradients along the line. The line is therefore divided into segments, or "divisions," which embrace gradients of a particular length or steepness. In order to ensure optimal utilisation of motive power over a particular division, the divisional boundaries must be drawn in such a manner as to concentrate all gradients of similar severity into a single division. The assembly of trains and allocation of 21 motive power takes place at the commencement and termination of each division, that is, at "divisional points." The amount of adjustment of train weight and motive power which is required at each divisional point is determined by the relationship between the severity of the gradients on the divisions adjacent to the divisional point. Minimum adjustment of train weight and motive power is commensurate with minimum cost for the operation over the contiguous divisions. Thus, "The grades of a division are very definitely related to each other in their effect on operation, and when considered in this connection may be termed the grade system of the division."5 The efficacy of the gradient system on the entire line is clearly dependent upon the efficacy of the gradient systems within the constituent divisions. The present analysis distinguishes between three types of gradient, the maximum gradient, the ruling gradient, and the pusher, or helper, gradient. The maximum gradient is simply the steepest gradient on the line. Operations over the maximum gradient may be conducted either by momentum, or by means of pusher locomotives, or by balancing traffic in such a manner that the heavier flow always descends the maximum gradient.6 The ruling gradient is that gradient which, "...by its length or steepness, limits the weight of train that can be hauled by_ one locomotive over the division on which it occurs."7 It should be noted that, The ruling gradient may or may not be the maximum gradient on a division. In the event that helper engines are used over the maximum grades, or momentum grades are employed...the next steepest grade becomes the ruling grade.8 22 The importance of the ruling gradient in the economics of railway operation is twofold. It determines not only the maximum train load which a single locomotive can haul over a particular division, but also the amount of motive power which is in effect "wasted" on those sections of the division which are not ruling gradients. Thus, ...it is not so much the direct cost of power that makes heavy ruling grades so objectionable, but rather the fact that this power which must be available wherever the ruling grade occurs cannot be used to advantage over other portions of the line.' Therefore, it is the ruling gradient rather than the maximum gradient which is the crucial determinant of operating expenses over a particular division. For it is the ruling gradient, its angle, length and frequency of incidence on the division, which together determine the amount of motive power which is provided over the entire division in excess of that required to move the train between the portions of ruling gradient. Alternatively, it is the ruling gradient which determines the amount of payload which must be left off the train over the entire division in order to ensure that a single locomotive can continue to haul the train whenever the ruling gradient occurs. The excess of motive power, or the loss of payload, may be minimised by the concentration of ruling gradients of similar severity within a particular division. Such concentration ensures that when a train is assembled to match the ruling gradient of the division, the motive power provided is fully utilised for as long as possible in crossing the division. A pusher or helper gradient is any gradient where an assisting locomotive is attached to the train in order to "help" 23 the train ascend the gradient. Pusher locomotives are required wherever the gradient of a section of the line exceeds the ruling gradient of the division.10 The assisting locomotives may be inserted into the train at any point along its length. However, the number of pushers which may be attached at any single point is limited by the strength of the drawbar of the car immediately before and the car immediately behind the pusher units, since it is upon these two drawbars alone that the full forces of buffing and pulling respectively are exerted. (The drawbar is that part of the locomotive or car which couples the vehicle to adjacent vehicles.) Moreover, with steam traction, as was operated over the Selkirk Mountains until the 1950's, co-ordination of locomotives in multiple is more difficult to achieve than with diesel and electric traction. Also, with steam traction, each additional locomotive requires its own train crew of at least two personnel, whilst diesel and electric locomotives may be operated in multiple by a single train crew. Pusher gradients, like ruling gradients, are expensive to operate, not simply because of the direct cost of supplying the additional power where it is needed, but because of the opportunity cost of being unable to utilise this additional power on portions of the line where it is not needed.11 Like ruling gradients, therefore, pusher gradients must be concentrated as much as possible. The steepness of the pusher gradient must also approximate as closely as possible the steepness of the maximum gradient negotiable by the number of locomotives in the train. In this manner, the maximum benefit is 24 derived from the "push," for if the pusher gradient is only slightly steeper than the ruling gradient, then much of the work of the pusher locomotives is not required in order to move the train.12, 13 Finally, the pusher gradient must be sufficiently long,14 and the traffic sufficiently dense,15 to ensure full utilisation of the pusher locomotive or the pusher fleet, i i) Curvature Curvature may be inserted into a line either in order to avoid a summit completely, or in order to moderate the gradient by which a summit is attained, or in order to avoid investment in a tunnel or cutting through the summit. In each case, the implications of the insertion of curvature for both construction and operation are identical. This analysis identifies two implications, those of resistance and distance, which are, however, interrelated in their impact upon constructional and operating costs. The resistance with which a train meets when travelling over straight track is increased when a curve is encountered, since the natural motion of the train is in a straight line. The increase in resistance due to curvature is identical in effect to the increase in resistance which results from an increase in gradient. Therefore, if curvature is inserted on the ruling gradient of a division, without reducing the gradient, the increase in resistance which this curvature entails is identical in effect to an increase in the ruling gradient: it necessitates either the attachment of additional motive power to each train which negotiates the curve, or the reduction of the payload of each train. 25 These expedients of additional motive power and payload reduction can only be obviated if the curvature is "compensated," that is, if the increase in resistance which the curvature entails is compensated by a decrease in the gradient, in order to ensure a constant level of resistance over both tangent and curvature.1' Thus, a gradient of "2.2 per cent, compensated" is a gradient on which the resistance encountered by a train is equivalent in amount to the resistance which the train would have encountered in ascending a 2.2 per cent, gradient on straight track. However, the actual angle of ascent will be less than 2.2 per cent., because some of the resistance encountered on the ascent will be due not to the gradient, but to the incidence of curvature. The angle of ascent is usually reduced, or "compensated," by 0.04 per cent, per degree of curvature.17 Thus, when a ten-degree curve is located on a gradient of 2.2 per cent., the actual angle of ascent over the curve must not exceed 1.8 per cent., for if it does, the train will stall, as the resistance due to the ascent, compounded by the resistance due to the curvature, will exceed the resistance of the ruling gradient of the division. The necessity to compensate gradients in order to avoid the stalling of trains implies increased distance, for either the angle of curvature must be reduced, or the angle of ascent must be decreased. The increase in distance which the insertion of curvature entails will increase either the immediate cost of construction or the delayed cost of operation. Each of these cost increases is particularly severe in mountainous terrain, where 26 construction costs per mile are initially high, and where very little local traffic can be generated by the lengthening of the line.18 Since the insertion of curvature may entail either increased distance at construction time or increased operating costs afterwards, it is not clear that curvature can always be regarded as a "Low capital cost, high operating cost" solution rather than a "High capital cost, low operating cost" solution to the problem of breaching mountainous terrain with rail. Its appropriate classification will always be determined by the specific circumstances in which it is adopted. The inference must be drawn that where curvature is adopted in practice, it represents the least-total-cost solution in comparison with either constructing a tunnel or operating over steep inclines. This analysis of "Low capital cost, high operating cost" solutions has demonstrated that the construction costs of overcoming mountains by rail, that is, the immediate cost of providing a railway facility through mountains, can only be reduced at the expense of incurring relatively higher operating costs once the facility has opened. These higher operating costs are incurred only when traffic is required to be transported over the facility. The costs are either direct costs consequent upon the deployment of more motive power, or opportunity costs consequent upon the foregoing of payload. It is appropriate to contrast the results of this analysis with an analysis of the obverse solution, that of "High capital cost, low operating cost." 27 I 2.1 (b) "High Capital Cost, Low Operating Cost" Solution. In the period of construction of a railway line through mountains, large sums of capital may be invested in order to eliminate an adverse gradient from the alignment by either tunnelling or cutting. The outlay is incurred "immediately," i that is, before any traffic can flow through the facility. ! However, once the facility has been completed, and the gradient eliminated, the cost of operating traffic over the line is less for every train than it would have been had the traffic been obliged to negotiate the steeper gradient, since the need to deploy more motive power or forego payload is averted. The high capital cost of gradient reduction represents a "lumpy" investment. Not only must the entire cost be borne "immediately," during the period of construction, but the investment project must be entirely completed before any savings in operating costs can be realised. Moreover, the rate of return on the investment is determined strictly by the volume of traffic which uses the facility. In the extreme case, if no I trains are run over the line, the fact that traffic can be moved at a low operating cost will be of no benefit in securing a return on the investment. Thus, the investment of large sums of capital per mile is not alone sufficient to ensure operating savings: traffic must be available in order to take advantage of the low operating costs. 28 2.2 The Trade-off The engineering and economic characteristics of alternative solutions to the problem of locating and operating railways in mountainous terrain require a trade-off between "construction costs" and "operating costs," and between "immediate costs" and "delayed costs." In order to determine the appropriate trade off, that is, in order to choose between alternative engineering schemes, the two questions posed in the Introduction have to be answered.19 First, what is the alignment which is appropriate to the traffic flows through the mountains? Where traffic flows are uncertain, or where there is an expectation that flows will be light, an alignment which entails high construction costs in order to obtain low operating costs will not be appropriate, for the benefit of the low operating costs will not accrue with sufficient frequency to offset the high interest charges on the capital invested in construction. The early North American transcontinental railways, built under uncertainty, or with the expectation of initially light traffic, therefore adhered to the "Low capital cost, high operating cost" solution: the first lines were built quickly and cheaply, in order to minimise interest charges for the future, and in order the sooner to generate revenues with which to upgrade the line and reduce operating costs as traffic developed. Immediate construction costs were diminished, and higher operating costs accepted in the short run.20 As long as the traffic volume remains low, the "Low capital cost, high operating cost" alignment continues to be appropriate. However, as the traffic volume increases, and as 29 operating costs escalate as a proportion of total costs, such an alignment becomes less appropriate, and the trade-off decision between capital costs and operating costs becomes less clear-cut. Ultimately, when the traffic volume increases to such an extent that the expected cost of operating trains over the alignment in the future exceeds the cost of constructing an alternative, less severe alignment, then it may be concluded that the "Low capital cost, high operating cost" alignment has become inappropriate for the volume of business which it is required to support. Capital must be invested in order to obtain a more appropriate alignment. In answering the second question, that is, what is the appropriate level of investment which should be undertaken in order to secure the appropriate alignment, it is instructive to consider the principle advanced in 1906 by the C.P.R. engineer who would later be charged with the task of locating a double track for the C.P.R. main line from Calgary to the West Coast: "The question of reducing grades over a certain section should be considered advantageous or economical when the saving effected in operating per annum over the section, due to grade reduction, more than represents the interest on the capital outlay necessary to make the reduction. . . "Figuring on the traffic being the same before and after the revision, the most economical location to make is the one which will require the least outlay for construction, and which will reduce operating expenses by an amount more than sufficient to pay interest on this outlay."21 This principle, which may be regarded as indicative of the C.P.R.'s own criteria for making the trade-off decision between construction and operating costs, and between immediate and delayed costs, is deceptively simple, especially when applied to the problem of investing in railways through mountainous 30 terrain, where the trade-off decision is rendered particularly difficult in practice by three factors. First, constructional and operating characteristics differ markedly between alternative alignments through mountains. This phenomenon in turn has two ramifications. First, the absolute level of investment required in order to obtain gradient reductions of any significance is likely to be high. Concomitant interest charges will also be high, therefore, and in order to offset these high interest charges, the possibility must exist of making large savings in operating costs. This possibility will exist only if the proposed gradient reduction project is drastic, or if the volume of traffic, forecast to benefit from the reduction in operating costs is considerable. Second, the range of alternative projects from which to choose is not likely to be "continuous." Thus, if the railway company is just unable to afford the investment in its "most preferred" alternative, the "next most preferred" alternative may offer significantly fewer benefits than the "most preferred" alternative in terms of construction costs and operating savings. The second factor which renders the trade-off decision difficult to make in practice is the necessity for the accurate estimation of construction costs, since the alignment should be adopted only if the operating savings relative to alternative investments will outweigh the cost of the initial investment. Clearly, the difficulty of accurately forecasting costs is a problem which pervades all project evaluations. However, the difficulty is particularly acute in the field of mountain railway location, where contingencies are 31 often difficult to foresee, and cost overruns easy to incur. The final factor which renders the trade-off decision difficult to make in practice is the necessity for the accurate estimation of operating savings. This in turn requires the accurate forecasting of future traffic volumes over the proposed alignment. Again, this is a difficulty common to all project appraisal. Again, too, however, the difficulty is particularly acute in the field of mountain railway location, where the scale of investment costs which must be recovered is usually large, and where flexibility to cope economically with extremes of traffic levels is difficult to incorporate into the proposed facility. The role of these three factors in shaping the trade-off decisions which were made by the C.P.R. through the Selkirk Mountains will be described and discussed in the remainder of this thesis. This analysis will discern and appraise the answers which the C.P.R., as revealed by their investment decisions, appear to have reached on the two questions posed above. The issue of "appropriateness," as elicited in response to each question, clearly involves more than simply engineering principles and economic costs and benefits. It involves the crucial matter of timing. When, and for how long, is an alignment to be considered "appropriate"? When should the investment be undertaken which is intended to secure a "more appropriate" alignment, and for how long is it intended that this alignment should be "more appropriate"? In analysing the nature and rationale of the trade-off decisions made in Rogers Pass, this thesis must perforce examine the manner in which the 32 C.P.R. addressed this crucial issue of timing. 33 FOOTNOTES 1 This discussion has no pretensions to be a comprehensive review of the engineering principles of mountain railway construction and operation, nor is it intended as such. For a thorough and contemporaneous treatment of the field of railway engineering, see, for example, A. M. Wellington, The Economic  Theory Of The Location Of Railways , New York, John Wiley & Sons, Inc., 6th edition, 1915; C. C. Williams, The Design Of Railway Location, New York, John Wiley & Sons, Inc., 1st ed., 1917; W. L. Webb, Railroad Construction, Theory And Practice, New York, John Wiley & Sons, inc., 8th edition, 1926. 2 Wellington, op. cit., Table 170, pp. 544-551. 3 See below, pp. 21-23. 4 "Considerably over half of the deterioration of track comes from the passage of engines over it, and the remainder only from the passage of cars, which may weigh ten or twenty times as much." Wellington, op. cit., p. 701. 5 Williams, op. cit., p. 219. 6 Ibid., p. 265. 7 Ibid., p. 219. My italics. 8 Ibid. ' Ibid., p. 220. 10 Unless the gradient is operated either by momentum or by balancing traffic. See note (6) above. 11 "It is a truth of the first importance, that the objection to high gradients is not the work which engines have to do on them, but it is the work which they do NOT do when they are thundering over the track with a light train behind them, from end to end of a division, in order that the needed power may be at hand at a few scattered points where alone it is needed." Wellington, op. cit., pp. 589-590. 12 Williams, op. cit., pp. 266-7. 13 "The rate of grade should be such as to require the full power of the pusher engine in addition to that of the regular engine to handle the maximum load over the balance of the section, as this will reduce the length of the pusher grade and consequently the pusher engine mileage." F. F. Busteed, "The Saving Effected By Grade Reductions," in, C.P.R. Co., "Proceedings Of The Meeting Of Western Lines Officials Held At Field, B.C., February twelfth and thirteenth nineteen hundred and six." Public Archives of British Columbia, Victoria, B.C. 34 (henceforth "PABC,") NWp 971B C225pr. p. 62. 14 "The maximum efficiency in operating pusher engines is obtained when the pusher engine is kept constantly at work, and this is facilitated when the pusher grade is as long as possible, that is, when the heavy grades and the great bulk of the difference of elevation to be surmounted is at one place. For example, a pusher grade of three miles followed by a comparatively level stretch of three miles and then by another pusher grade of two miles cannot all be operated as cheaply as a continuous pusher grade of five miles." Webb, op. cit., pp. 580-81. 15 "...the condition that the pushers must be kept busy and be always on hand to have them economical must be remembered. The larger the traffic of the road the more easily can this be assured, and consequently the more frequently can pushers be used." Wellington, op. cit., p. 606. 16 Williams, op. cit., p. 296. 17 Webb, op. cit., p. 563. 18 It should be noted that compensation is also required in railway tunnels which are located on gradients. Here, the increase in resistance is due to damp rails and increased air resistance within the tunnel. 19 See above, p. 3. 20 "Whereas European engineers inclined to a permanent type of construction, American railroads were often best built when most cheaply built, with light rails, sharp curves and steep grades. Only such roads could expect to earn interest on their investment, in view of the scant population of the country and the pioneer character of many of the early enterprises." S. Daggett, Principles Of Inland Transportation, New York, Harper, 4th edition, 1928, pp. 63-64. 21 F. F. Busteed, op. cit., p. 62. 35 CHAPTER 3 RAILWAYS AND ROGERS PASS The purpose of this chapter is to relate the engineering and economic concepts of the previous chapter to the specific circumstances of railway construction and operation in Rogers Pass, B.C. The chapter is divided into three sections. The first section describes the physical and climatic characteristics of the Selkirk Mountains, and considers the implications of these characteristics for railway location and operation. The second section explains why Rogers Pass was selected from the alternative routes available for the location of the transcontinental main line through the mountains of B.C. The third section examines the specific expectations which the C.P.R. harboured for the impact of these characteristics upon prospective constructional and operating conditions in the Pass. 3.1 Rogers Pass The direct route west from the foot of the Kicking Horse Pass crosses the northerly flowing Columbia River, and is then faced by the great mass of the Selkirk Mountains (See Map I). These mountains pose a number of problems for the construction and operation of railways. These problems are related to the location, the topography and the climatic characteristics of the Selkirk Mountains. 37 a) Location The Selkirk Mountains form a chain lying to the west of the Rocky Mountains. They are divided from them by the Columbia Valley, running approximately north and south, and through which the river of the same name flows. This river sweeps round the northern extremity of the Selkirk chain, forming what is called the 'Big Bend,' and then flows southerly into Oregon Territory, scooping out a deep valley, which divides the Selkirks from the Gold Range, lying further to the west. The Selkirks are thus bounded on either side, and enclosed at their northern end, by the Columbia Valley. Their length is about 250 miles in Canadian Territory, and width from 50 to 80 miles.1 Situated just inside, and parallel with, the eastern border of B.C., the Selkirks are the second of four great mountain ranges, the Rockies, the Selkirks, the Gold Range and the Coast Range, which stand astride southern Canadian routes from the prairies to the west coast. b) Topography In general character (the Selkirks) are lofty, rugged, and steep; intersected and diversified by narrow passes, and precipitous, rocky canons [sic]. The height of the highest peaks is ten or eleven thousand feet above the sea; long parallel ridges of not much inferior elevation may be frequently observed in close proximity, forming between them a narrow V shaped valley, whose sides extend upwards, at an even and very steep slope, for five or six thousand feet, and along the bottom of which there flows a turbulent mountain stream.2 A "low capital cost" railway location, intending to follow the natural alignment of the terrain,3 would necessarily seek these narrow passes as corridors through the mountains. However, the valley floors in the Selkirks are poorly drained, marshy, and densely overgrown.* Even today, the Trans - Canada Highway is compelled to skirt these areas by taking to the valley sides. Thus, even where level valley floors exist, and where they are wide enough to accommodate even a single line of railway, no 38 track can be laid without massive investment in ground clearance and drainage. Instead, track must be thrown up the valley sides, clear of the area prone to flooding. Moreover, where the valley floors are drained, the streams descend very steeply towards the major rivers, too steeply to be followed by railway lines. In seeking less precipitous descents, tracks necessarily diverge from the river valleys and are thrown high up the valley sides.5 They must then be "developed" in order to reach the valley floors by practicable gradients, thereby incurring increased distance and curvature. The necessity of abandoning the valley floors and seeking the mountain sides, whether induced by the narrowness of the valley, poor drainage or steeply flowing rivers, has six consequences adverse to railway construction and operation. First, it entails steep gradients, in order to reach the refuge of the valley sides and then return to the valley floor where possible. Second, it involves construction through the densely forested valley sides, which would be scarcely less expensive to clear than the dense undergrowth of the marshy valley floors. Third, it entails extensive cutting into the side of the mountain above the valley floor, in order to carve out a "bench" upon which to lay the rails. The extent of this cutting is increased by "development" of the line. Fourth, it involves increased expenditure upon the securing of a stable foundation for the line over these "benches," since in the Selkirks, "the rock being for the most part...clay and slate shales...crumbles and degrades easily under the action of the weather, and large masses of debris are thus constantly gathering in the valley 39 bottoms, while the mountain sides are deeply scarred by gullies and fissures."6 Fifth, it involves bridging these gullies and the many mountain streams, all of which are avoided in an alignment along the valley floor.7 The bridges may be short, but the crossings will be high above the gullies and streams, and the bridging will therefore be expensive and complex. Moreover, allowance must be made, in both the length and the strength of the bridges, for the violent flooding of these many streams. This flooding, the result of warm weather melting "the snow-fields and ice masses..., may occur at any period of the summer months, and may last for days, or perhaps weeks."8 Finally, location along the mountain side leads the tracks perpendicularly across the paths of avalanches descending from the peaks above to the valley floor beneath. This problem of avalanches will be examined further in the consideration of climatic characteristics below. c) Climate The Selkirk chain forms, as it were, a lofty wall running north and south. Being very much higher than the mountains to the west, it is the first and chief barrier that the moisture laden currents of air from the Pacific Ocean encounter on their eastward passage. This warm air is intercepted and the moisture condensed by contact with the cold Selkirks, entailing heavy rain in summer and deep snow in winter...' The average annual snowfall ranges from thirty to fifty feet,10 falling mostly between October and April,11 and falling much more heavily upon the western slope than upon the eastern slope.12 The heavy snowfall, coupled with high winds and the steep, fissured profile of the mountains, creates a severe avalanche danger. The severity of the danger lies in the 40 velocity of the avalanches, the volume of snow which slides, and the weight of that snow. Hard-packed and frozen, the snow alone may weigh from 25 to 38 lbs. per cubic foot, and the force of the slides may tear down whole trees or rocks and carry them into the valleys with the avalanche.13 The time of year when slides are largest and most frequent is from the middle of January to the latter part of February. These are 'winter slides,' formed of large masses of quite dry snow. In March and April there are numerous 'sun slides,' caused by the melting of the snow and ice, but these are not of any importance as compared with the others.14 The incidence of these slides poses a considerable seasonal hazard to both the construction and the operation of transportation corridors through the Selkirks, and greatly increases the expense of keeping open those corridors. It has already been noted that in being compelled to abandon valley floors, railway lines would be forced directly across the paths of avalanches on the mountain sides. However, even the valley floors do not necessarily provide a refuge from the avalanches. Many of the valleys are too narrow to permit the slides to "run off" harmlessly, and the larger avalanches acquire sufficient momentum to cross the valley floors and travel considerable distance up the sides of the mountains opposite. Structures intended to defend railway lines from the slides must be strong enough to withstand not merely the weight of the cumulative snowfall, nor even the strength of the avalanches themselves, but also the weight of the falling rocks and trees which accompany the slides. Clearance of the snow slides alone would be arduous and expensive enough: many of them would cover several hundred feet of line, sometimes to a depth of over 41 thirty feet.15 The presence of rocks and trees in the debris exacerbates the clearance problem, particularly if mechanical ploughs are used in the clearing operation. This introduction to the geography of the Selkirk Mountains has highlighted the physical constraints upon railway location and operation in those mountains. The constraints are imposed by the topography and the climate of the Selkirks. It should be noted that these constraints are to a considerable extent peculiar to the Selkirk Mountains. In the Rockies, the valley floors are generally wider, better drained and less overgrown, and are thus eminently suitable to accommodate transportation corridors. Moreover, being situated further east than the Selkirks, they receive a far lesser snowfall, mitigating the problems of avalanches and snow clearance. Even in the Sierra Nevada, which experiences considerable snowfall and avalanches, the problems are far less severe than in the Selkirks, for the snow itself is much lighter, and the snow slides do not generally displace rocks and trees.16 Moreover, the constraints are to a considerable extent peculiar to the location and operation of railways. Due to the superior adhesion, acceleration and cornering capabilities of road vehicles, highways may follow the severe gradients of the mountain streams, or may alternatively choose both severe gradients and curvature as means of avoiding avalanche paths, gullies or flooded valleys. These alternatives are available to rail in only limited measure. This analysis of the implications of the physical and climatic characteristics of the Selkirk Mountains for railway 42 construction and operation has revealed that the nature of the terrain would pose a formidable challenge to the location of a rail corridor from the prairies to the west coast. One may legitimately wonder why and how a transcontinental rail link could be expected to penetrate this awesome natural barrier against communication. The next two sections of this chapter will endeavour to answer these questions. 3.2 The Select ion Of Rogers Pass For The First Transcontinental  Rail Link The purpose of this section is to explain why Rogers Pass was selected as one segment of the route by which the C.P.R. would cross the mountains which stood between the prairies and the west coast. As a result of this decision, the C.P.R. would be for ever implicated in the struggle with the adverse physical and climatic characteristics outlined above. It should be noted that the selection of Rogers Pass for the transcontinental link was the consequence of two separate but interrelated decisions. The first decision addressed the question of the general alignment which should be adopted in crossing the western Canadian mountains. The second decision addressed the question of the specific location which should be adopted in crossing the Selkirk Mountains of B.C. This analysis will concentrate upon the second of these questions, partly because the first question has been discussed thoroughly elsewhere,11 and partly because the second question reveals more than the first question about the manner in which the trade-offs elaborated in Chapter 2 were handled in the particular environment of the Selkirk Mountains. 43 The need to cross the Selkirk Mountains could have been avoided entirely had the C.P.R. adhered to their original contract and constructed the main line through the Yellowhead Pass via Jasper House. Such a route would have had maximum gradients of one per cent., and would have been free from snow slides.18 However, shortly after the C.P.R. had been awarded the contract to build the transcontinental main line, in 1881, a search was initiated for a more southerly route across the prairies and through the mountains. By 1883, the Yellowhead Pass alternative had been abandoned. It is not proposed to reopen the controversy which surrounds the rationale for this abandonment decision. The objective of the C.P.R. in seeking a more southerly route was ostensibly to obtain a shorter line.15 This saving in distance was equated by the C.P.R. with a reduction in future operating costs,20 and with an improvement in their capability to compete with American rivals for transcontinental traffic.21 In order to secure these savings, it appears that the C.P.R. was prepared to sanction higher construction costs on a shorter line through the mountains.22 However, the decision to abandon the Yellowhead Pass may also have been motivated by political considerations, since a route further south would more readily forestall the economic encroachment of the United States into both the Canadian prairies and southern B.C.23 Certainly, the decision to reject an extreme southern crossing of the mountains appears to have been dominated by political considerations. Sir Thomas Shaughnessy, third President of the C.P.R., would later assert that the Company 44 "would have preferred to build via the Crow's Nest, but the Government of the day did not approve this, as the line would be too close to the International Boundary. As a consequence the present route was adopted."24 That "present route" would penetrate the Rockies through the Kicking Horse Pass and the Selkirks through Rogers Pass. It was the decision to follow the Kicking Horse Pass which made necessary a decision concerning the appropriate crossing of the Selkirk Mountains. Two alternatives were available. An alignment could be followed around the Big Bend of the Columbia Valley, skirting the northern extremity of the range, or, alternatively, a direct crossing could be sought through the Selkirks. The C.P.R.'s evaluation of these alternatives will now be assessed. From this assessment, it will be possible to analyse the manner in which the C.P.R. handled the trade-offs explained in Chapter 2. The analysis will be instructive for the appraisal of later trade-off decisions made by the C.P.R., because the trade-offs made in the Selkirks appear not to have been coloured by political considerations, but to have been based purely upon principles of transportation economics. The length of the Big Bend route was estimated at 140 miles, and it seemed "quite certain that gradients of 80 or 90 feet per mile would have to be used in places."25 Since the ruling gradient elsewhere on the C.P.R. system was to be 52.8 feet per mile,26 these sections of the Big Bend route would have to be operated as pusher gradients. Moreover, since the gradients would be short and dispersed, the pusher operation would be difficult to conduct economically. Finally, the 45 adoption of the lengthy alignment around the Big Bend would not necessarily preclude the need for tunnelling, although no estimate has survived of the actual length of tunnelling which might have been required.27 When the C.P.R. applied for statutory authority to abandon the Yellowhead Pass, the Big Bend represented a "fail-safe" alternative on the southern route: it could be adopted as a last resort if no direct crossing of the Selkirks could be found.28 However, if the C.P.R. had been driven to adopt the Big Bend, the potential saving in distance on the southern route, which had induced them to abandon the Yellowhead Pass, would have been eroded to between thirty-five and forty-five miles.25 It is doubtful that the savings in the cost of operating this distance, particularly over the potentially uneconomical pusher gradients, would have offset the increased construction costs which were anticipated in the Kicking Horse Pass.30 If adoption of the southern route were to be justified on economic grounds, therefore, the C.P.R. had to be prepared to invest even more heavily at construction time in order to ensure that the distance and cost of operating around the Big Bend would be saved. As the C.P.R.'s General Manager, Van Home, freely admitted, "'to save this distance work will be undertaken that would ordinarily be considered impracticable on account of expense.'"31 It was therefore understood and expected that a direct crossing of the Selkirks would necessitate a heavy "immediate" investment. The exact amount which the C.P.R. was prepared to invest in order to secure the saving in distance is not known, 46 but an interesting revealed-preference function may be deduced, which offers at least a general indication of the extent of that preparedness. The engineer Walter Moberly, 'having discovered Eagle Pass through the Gold Range in the 1860's, sought to link it with a pass through the Selkirks, but in 1871 he had abandoned his surveys with the conclusion that such a pass "would be impracticable for a railway unless a long tunnel, probably 14 to 15 miles in length, should be excavated through the Selkirk range."32 The report of the necessity for this length of tunnelling probably expedited the Federal Government's decision in favour of the Yellowhead Pass, announced while Moberly was in Victoria, having returned from his 1871 explorations.33 Ten years later, when the C.P.R. requested permission to adopt a southern route, the expectation was still that "some long tunnels" would be required in order to secure a direct crossing of the Selkirks.34 Although, as has been noted above, the C.P.R. was prepared to undertake "work... which would ordinarily be considered impracticable on account of expense," it is not clear that they regarded this amount of tunnelling as acceptable, for they continued to reserve the option of building around the Big Bend.35 The amount of tunnelling which the C.P.R. was prepared to accept appears to have been a maximum of 2 1/2 miles, for it was only at the end of the 1882 surveying season, after Major A. B. Rogers had established that no more tunnelling than this would be required if a route through Rogers Pass were to be adopted, that the C.P.R. applied specifically for permission to locate 47 the line directly across the Selkirks.36 When even this amount of tunnelling proved not to be necessary in order to secure acceptable gradients, the direct crossing via Rogers Pass became clearly preferable to the circuitous passage around the Big Bend, saving some seventy-seven or eighty-seven miles of mountain railway construction.37 The sketching of this preference function has offered some insight into the extent to which the C.P.R. was prepared to undertake "immediate" investment in order to secure future operating savings. It is necessary now to establish the extent of the operating savings which the C.P.R. would require in return for its willingness to accept higher construction costs. Again, no quantitative data is available. Again, too, however, it is possible to obtain some insight into the trade-off from the evidence of preferences, both revealed and explicit. Evidence of revealed preference is available in the C.P,R.'s preparedness to accept gradients steeper than initially projected on the Rogers Pass route. Rogers, when surveying the Selkirks in 1881, had initially reported "a grade not to exceed sixty-six feet to the mile between Kamloops and the North Fork of the Illi-cille-want [sic], and from thence to the summit of the Selkirks not to exceed eighty feet to the mile."38 After the following year's surveys, he was compelled to revise these estimates upwards, locating "a line ascending westerly for a distance of twenty miles to the summit of the Selkirks at the rate of 105 6/10 feet per mile, and descending the western slope at the same rate for the same distance..."3' Nevertheless, the C.P.R. was clearly aware of this gradient system when they 48 applied for permission to exploit Rogers Pass, and was clearly prepared to accept it.40 When Rogers later recommended that the gradients again be revised upwards, to 116 feet per mile,41 it was still not expected that the revision would negate the advantages of Rogers Pass over the Big Bend. "(I)nasmuch as assistant engines would be required on a grade of ninety feet as well as on one of 116 feet per mile...,"42 better utilisation could be anticipated from pushers over the Selkirk summit than from pushers on scattered gradients around the Big Bend. Therefore, the savings in distance would not be offset by the costs of having to operate over the steeper gradients. Evidence of explicit preference is available from Van Home's own retrospective explanation of the trade-off decision. In his evaluation, Van Home assumed that no pusher gradients at all would be required on the Big Bend. Therefore, the savings in distance over Rogers Pass would be offset by the full cost of the pusher operation which would be required on the summit route. Even under this assumption, however, the Rogers Pass alternative was still preferred. Explicitly, the anticipated savings in distance would offset the cost of operating over steeper gradients.43 Implicitly, from the above evidence of revealed preference concerning construction costs, it may be inferred that the savings in distance were also expected to offset the potentially higher cost of construction through Rogers Pass. From the evidence of both revealed and explicit preference, it is clear that the prime factor in influencing the decision to 49 adopt Rogers Pass was the anticipated saving in operating costs, and in particular, in the operating cost of distance. Insofar as this saving in the operating cost of distance was equated with an improved capability to compete for through traffic, the manner in which the trade-off decision was handled in the Selkirk Mountains was consistent with the stated objectives which had prompted the search for a more southerly location.44 This analysis explains why the main line of the first Canadian transcontinental railway was located through Rogers Pass. The decision to dismiss the Yellowhead Pass and Crow's Nest Pass alternatives was based at least partly upon political grounds. Nevertheless, it is by no means clear that this decision was inconsistent with a decision which would have been based purely upon principles of railway economics. With the selection of Rogers Pass, it appeared that the objectives which the more southerly location was intended to achieve had been fulfilled, and that the trade-off which the C.P.R. had been prepared to make, that of incurring higher construction costs immediately in order to save operating costs later, had been made. Indeed, when Rogers Pass was selected from the alternatives, it was not at all clear that construction costs would be higher in building through the Pass than they would have been in building around the Big Bend, for although the C.P.R. had been prepared to accept 2 1/2 miles of tunnelling if it had had to, in practice there were virtually "no tunnels necessary."45 Moreover, although the C.P.R. had been prepared to accept that the cost of the pusher operation over Rogers Pass would offset some of the saving in distance, in practice it is 50 likely that the cost of the pusher operation over Rogers Pass was actually less than the cost would have been of the pusher operation which would certainly have been required around the Big Bend. The selection of Rogers Pass, therefore, was not merely a necessary expedient. It was not an alternative forced upon the C.P.R. by the Company's blind decision to enter the Rockies via the Kicking Horse Pass with no sure knowledge of its means of exit.*' Rather, it was a positive choice. It represented a sound solution to the problem of crossing the Selkirk Mountains by rail. The C.P.R. was well pleased,47 and expectations were high. It. is appropriate now to consider the nature of these expectations, and the foundations for optimism. 3.3 The Expectations Of The Builders The purpose of this section is to examine further the expectations which the C.P.R. harboured for construction and operating conditions over Rogers Pass, prior to the actual construction and inception of the transcontinental link. The examination highlights the gap between the C.P.R.'s expectations of the route, and the realities which will be examined in the following chapter. The analysis of the C.P.R.'s specific expectations in Rogers Pass will examine five distinct areas of constructional and operational concern. These areas are the character of construction work, the cost of construction, the time required for' construction, the methods of operation, including forecasts of traffic volumes, and the necessity for protection of the facility from snow slides. 51 a) The Character Of Construction Work On first traversing the Pass in 1882, Rogers himself had felt "entirely safe in reporting a practicable line through this range," although he expected that the work would be "very heavy and expensive."48 The following year, after further surveys, Rogers again reported optimistically: "Through the Selkirks the work is more uniformly distributed than through the Rockies and presents no special engineering difficulties and for mountain work may be considered moderate, the percentage of rock being unusually small."45 Tunnelling was expected not to exceed 1,200 lineal feet on the entire distance across the Selkirks, in comparison with 1,800 feet on the Upper Kicking Horse, 1,400 feet on the Lower Kicking Horse and 2,200 feet in the Columbia Canyon.50 In a "Memorandum of the General Character of the Work," prepared in February 1884, the C.P.R.'s Chief Engineer observed that, From the east foot of Selkirks to mouth of Eagle Pass:-The work may be considered moderate for mountain work, being largely composed of gravel."51 Again this contrasted with conditions in the Rockies, where, on the west slope, in the Chief Engineer's estimation, "The work may be classed as generally heavy, with some short distances very heavy."52 In September 1884, mere months before construction through the Selkirks commenced,53 S. B. Reed, the engineer who had located the mountain section of the Union Pacific Railroad, reported that, The line over the Selkirk Mountains, a distance of sixty-three miles, is remarkably easy to construct, there being comparatively little rock excavation, and but one short tunnel. The great bulk of the work will be in earth and loose rock.54 52 b) The Cost Of Construction Few specific forecasts of construction costs in the Selkirks survive.55 In February 1884, the House of Commons of Canada had been informed that the C.P.R.'s estimate for the entire distance from the summit of the Rockies to Kamloops was $12 million,56 an average of $44,776 for each of the 268 miles. When Reed travelled the route in August 1884, he calculated that, "between the summit of the Gold Range and the summit of the Rocky Mountains... this section of the road can be constructed at an average cost not exceeding thirty three thousand dollars ($33,000) per mile."57 Reed's glowing conclusion was that, In view of the rugged mountain country, through which the line passes, from Savonna [sic] Ferry to the summit on the main range of Rocky Mountains, a distance of two hundred and ninety miles...you have an exceedingly cheap line to build, costing far less per mile than the mountain work of the Union and Central Pacific roads."5" The more conservative estimates which the C.P.R. formulated for submission to the Federal Government the next month forecast an average cost for this same section of some $37,000 per mile, and the forecast cost of construction across the Selkirks was actually slightly below this average.59 Even if it is assumed that the Big Bend could have been operated as cheaply as the direct crossing, construction costs around the Columbia River would have had to have been less than $20,000 per mile for the Big Bend route to have represented a cheaper overall solution than the Rogers Pass route to the problem of crossing the Selkirk Mountains by rail.60 53 c) Time Required For Construction In their original contract with the Federal Government, the C.P.R. had been committed to complete the entire transcontinental facility within ten years, that is, by May 1891. Rapid progress across the prairies was doubtless a major factor in enabling this deadline to be brought forwards, but the location of a direct route through the Selkirks, and the optimistic projections of its engineering feasibility, must also have contributed to the revision of the target. When the C.P.R.'s President, Sir George Stephen, reported the discovery of Rogers Pass to the Marquis of Lome in September 1882, he volunteered: "Expect to have the whole line from Montreal to Pacific Ocean open by January first, 1887,"61 over four years sooner than the contracted deadline. By December 1882, even this expectation had been revised. Stephen stated that the C.P.R. expected "to complete their own work across the mountains" during 1885.62 Up to this time, the Lake Superior section was expected to be the last completed,, being scheduled to open during 1886.63 These expectations were unchanged a year later, when the C.P.R. concluded a contract for financial assistance from the North American Railway Contracting Company which stipulated completion of the mountain section by December 31, 1885, and completion of the Lake Superior section by December 31, 1886.64 When this contract lapsed, and the Federal Government again intervened with loans to the C.P.R., the expectations were revised. The C.P.R. undertook completion of the entire project by May 1886, thus buying time in the mountains at the expense of time on Lake Superior.45 In May 54 1884, Van Home confided to Major Rogers, "(W)e hope that the men on construction from the East will reach the second crossing of the Columbia and possibly Eagle pass by the end of this year..."6' Slow progress down the western slope of the Rockies67 only slightly tempered this confidence. In September 1884, Van Home assured the Directors, "I think there will be no difficulty in completing the mountain section within a year from this date.. d) Operating Methods And Traffic Forecasts As the above analysis demonstrated,69 the C.P.R. was aware when Rogers Pass was discovered that whether the main line followed the Big Bend or traversed the Pass, a pusher operation would be required for either alternative. Several arguments were advanced in favour of the pusher operation over Rogers Pass. The pusher gradients would be concentrated within twenty miles on either side of the Selkirk summit,70 permitting intensive utilisation of the pusher fleet. The summit itself was "represented as being admirably adapted for the location of a depot for marshalling trains, being practically level for a distance of about three quarters of a mile."71 Moreover, "considering the fact that the heavy grades in the Selkirk Range are embraced within a comparatively short distance, their disadvantage is very little as compared with the great savings in through distance."72 Less pusher capacity would be wasted over Rogers Pass than on the lighter pusher gradients around the Big Bend.73 Finally, since the only other pusher gradient was expected to be for twenty miles on the west slope of the 55 Rockies,74 the pusher gradients in the Selkirks complemented the gradient system of the entire transcontinental railway, a system which compared favourably with those of the Central and Union Pacific Railways, the standards of which had provided the model for the C.P.R.7 5 It is likely that these arguments alone would have sufficed to persuade both the C.P.R. Directors and the Federal Government that the pusher operation over Rogers Pass would not be detrimental to the movement of traffic through the mountains of B.C., but rather, in conjunction with the saving in distance, would be a strength of the southern route. Two further arguments in favour of the pusher operation were presented by the C.P.R. The first was that traffic requiring transit through the mountains would be light "for a number of years to come."76 Therefore, a small fleet of pusher locomotives would suffice to handle the business. In the Rockies, Van Home forecast that "three, or at most four, trains each way per day will carry all the business to be done...,"77 and expressed the belief "that in the case of passenger trains double locomotive service will seldom be required; ordinarily the substitution of a heavy for a light locomotive will answer the purpose."78 The forecasts of traffic volume appear internally inconsistent with the boasts of timber and mineral resources vaunted by both Stephen79 and Van Home.80 Nevertheless, Van Home was confident that such traffic as was carried through the mountains would be carried profitably.81 The second argument was "that the preponderance of through traffic across the continent (would be) largely west bound, and I 56 that the two heavy gradients rising eastward might therefore be still heavier without material disadvantage."82 This forecast flatly contradicted those generated by members of the B.C. Board of Trade, which emphasised eastbound flows from B.C.83 However, the argument does highlight the "system" implications of the pusher gradients over the Selkirks. On the entire C.P.R. transcontinental "system," a westbound preponderance of traffic was preferable, since this was revenue traffic, and the trains conveying revenue traffic would have to be tailored to only one restrictive ruling gradient, while longer trains of empties, generating zero direct revenue, could be hauled against the two adverse eastbound ruling gradients.84 Thus, the.traffic imbalance would actually reduce the total joint cost of the operation. Within the Selkirk Mountains themselves, however, since the pusher gradients were expected to be of approximately equal length, a perfect balance of flows would be optimal, in order to ensure even utilisation of motive power on either side of the summit. An imbalance in favour of either direction would be equally costly, but the actual direction of the imbalance would be irrelevant to the economics of the pusher operation. e) Snowslide Protection The C.P.R. had long been aware of potential avalanche problems in crossing the Selkirks directly.85 They had accepted Rogers' recommendation that gradients through Rogers Pass be increased from 105.6 feet per mile to 116 feet per mile, "in order to avoid some points where dangerous snow slides are to be feared."8' This recommendation had been relatively "cheap" to 57 implement,87 but further gradient increases in order to avoid slides were impracticable, since they would entail a costly increase in the ruling gradient of the system, and breach of their original contract, which had stipulated maximum gradients of 2.2 per cent, compensated. Tunnelling in order to avoid avalanches does not appear to have been considered at construction time.88 Instead, where avalanche paths could not be avoided, the C.P.R. intended to protect the main line with snow sheds. When Rogers had first traversed the Pass, he had felt "assured that the distance in which difficulties may be expected in crossing the Selkirk Range will be reduced to ten or twelve miles."85 As late as August 1884, Reed, making the same crossing, would report that "evidences of snow slides were seen at and near Roger's [sic] Pass, in the Selkirk Range, also near the summit of the main range of the Rocky Mountains, but the aggregate distance on which these occur does not exceed fifteen miles."50 He admitted that, "A number of snowsheds will probably be required for the protection of the track," but pointed out that "nearly fifty miles of these are in successful use on the Central Pacific road."51 Thirty-two miles of these sheds had cost $1,731,000 in the 1860's.52 In March 1885, the C.P.R. estimated that $450,000 would be required for the construction of snowsheds in the mountains53: if they were expecting to be able to construct their sheds for the same cost as the Central Pacific had incurred, they could expect to completely cover with snowsheds at least eight miles of those fifteen troublesome miles identified in the mountains. This detailed investigation of the specific expectations 58 which the C.P.R. harboured for construction and operating conditions over Rogers Pass prior to the actual commencement of work in the Selkirks reveals the particular grounds upon which rested the Company's satisfaction in securing a location through Rogers Pass and their confidence in contemplating future operations over the Selkirks. Construction was expected to be relatively easy and significantly less costly than alternative routes. Adoption of the Rogers Pass route would enable early completion of the entire transcontinental facility, and the rapid generation of revenues from through traffic with which to support subsequent improvements to the line. The pusher operation over the Selkirk summit would complement the gradient system of the entire transcontinental facility, and it was expected that the avalanche problem, which did not appear unduly burdensome when set against conditions experienced by rival railways, would be effectively eliminated by a modest capital outlay at construction time. The results of this investigation reinforce the conclusion reached in the second section of this chapter, that the Rogers Pass route appeared to offer a positively sound rather than a merely expedient solution to the problem of breaching the Selkirk Mountains by rail. The C.P.R. had been prepared to undertake heavy "immediate" investment in order to obtain operating savings in the future. Yet not only did the operating costs over Rogers Pass appear likely to be far less than operating costs around the Big Bend, but it seemed that a heavy "immediate" investment of capital would not be necessary in order to obtain the savings in operating costs. Thus, although 59 the C.P.R. had been prepared to make a trade-off between construction costs and operating costs, it seemed that they would not in practice have to make such a trade-off, for the Rogers Pass route represented the least-capital-cost and least-operating-cost solution. Moreover, the decision to adopt Rogers Pass in preference to the Big Bend was much less controversial than the decision to adopt the Kicking Horse Pass in favour of the Howse Pass through the Rockies. The engineers Fleming, Hogg, Rogers and James Ross were each independently dispatched through the Howse Pass in various attempts to establish the feasibility of that alternative. It was not until November 1883, after the railhead had already advanced far to the west of Calgary, and after even more surveys through Howse Pass had been undertaken, that Ross, the manager of construction in the mountains, would admit to feeling "quite satisfied that we have secured beyond doubt the best line through the Mountains."'4 The wrangling with the Ministry of Railways and Canals which accompanied the submission of the C.P.R.'s location plans for the western slope of the Rockies'5 and the Kamloops Lake sections'6 was completely absent from the submission of the profiles of the alignment over Rogers Pass. These latter plans were approved and returned quietly, quickly and without question during the autumn of 1884,57 in ample time for the commencement of construction through the Selkirks the following spring. 60 FOOTNOTES 1 G. C. Cunningham, "Snow Slides in the Selkirk Mountains," Transactions of the Canadian Society of Civil Engineers, Vol. I, Part II, October-December 1887, p. 18. 2 Ibid., pp. 18-19. 3 See above, pp. 17-18. * For a graphic description of the difficulties of penetrating the Selkirks on foot, see, S. Fleming, England and Canada. A Summer Tour between Old and New Westminster, with Historical  Notes, Montreal, Dawson Brothers, 1884, pp. 271-94. In three full days of marching, Fleming's party managed barely ten miles through the Selkirks. 5 Cunningham, op. cit., p. 19. * Ibid. 7 Ibid. 8 W. S. Vaux, Jr., "The Canadian Pacific Railway from Laggan to Revelstoke, B.C.," Reprinted from the Proceedings of the  Engineers' Club of Philadelphia, Vol. XVII, No. 2, May 1900, p. 72. 5 Cunningham, op. cit., pp. 19-20. 10 A. C. Dennis, "Construction Methods for Rogers Pass Tunnel," Proceedings of the American Society of Civil Engineers, Vol. XLIII, No. 1, January 1917, p. 6. 11 T. C. Keefer, "The Canadian Pacific Railway," Transact ions of the American Soc iety of Civil Engineers, Vol. XIX, No. 394, June 1888, pp. 83-84. 12 Cunningham, op. cit., p. 20. 13 Ibid., pp. 20-24. 14 Ibid., p. 24. 15 For an analysis of the frequency and mass of avalanches on major avalanche paths in the Selkirk Mountains from 1909 to 1979, see, B. B. Fitzharris and P. A. Schaerer, The Frequency of  Major Avalanche Winters, Ottawa, National Research Council Of Canada, Division of Building Research, June 1979. Cunningham himself recorded one snow slide standing forty feet deep on the roof of a C.P.R. snow shed. Cunningham, op. cit., p. 30. 16For an account of snow problems in the Sierra Nevada, see, G. M. Best, Snowplow: Clearing Mountain Ra i1s, Berkeley, 61 California, Howell-North Books, 1966. See also Cunningham, op.cit.,p.25. 17 See, for example, W. Kaye Lamb, The History of the Canadian  Pacific Railway, Toronto, Macmillan, 1977, pp. 79-81; J. H. E. Secretan, Canada's Great Highway: From the First Stake to the  Last Spike, London, John Lane, 1924, pp. 247-8; C. A. Shaw, op. cit., pp. 10-11; R. G. MacBeth, The Romance of the Canadian  Pacific Railway, Toronto, The Ryerson Press, 1924, p. 85; N. Thompson and J. H. Edgar, Canadian Railroad Development from the  Earliest Times, Toronto, The Macmillan Company of Canada, 1933, p. 138; M. Sprague, The Great Gates; the story of the Rocky  Mountain passes, Boston, Little, Brown, 1964, p. 293; G. P. de T. Glazebrook, A History of Transportation in Canada, Toronto, The Ryerson Press, 1938, p. 275; A. J. Johnson, "The Canadian Pacific Railway and British Columbia, 1871-1886," MA Thesis, University of British Columbia, 1936, pp. 151-155. 18 E. E. Pugsley, The Great Kicking Horse Blunder, Vancouver, 1973, pp. v-vi. 15 A C.P.R. engineer in the west would recall the day that, "Van Home sent for me, and announced in a most autocratic manner that he wanted "The shortest possible commercial line" between Winnipeg and Vancouver..." Secretan, op. cit., p. 99. 20 "...the Canadian Pacific Railway company propose to carry their railway far to the South of Edmonton if a practicable line can be found by the Kicking Horse Pass that will shorten the distance very considerably and thereby reduce the cost of operating it." Marcus Smith to Collingwood Schreiber, Chief Engineer of the C.P.R., April 10, 1882, Department of Railways and Canals, Railway Branch, Central Registry Files, Public Archives of Canada, Ottawa, (henceforth "PAC"). RG 43 A 2 (a) 6710 Vol. 223. 21 "The importance of the great saving in distance by this line cannot be overestimated. It affords a line across the continent materially shorter than that from New York to San Francisco by way of the Union and Central Pacific Railways, and places beyond a doubt the ability of this Company to compete successfully for the trans-continental freight and passenger traffic." Charles Drinkwater, C.P.R. Co.' Secretary, to Sir Charles Tupper, Minister of Railways and Canals, February 21, 1883, Dominion  Sessional Papers, Ottawa, (henceforth "DSP,") Vol. XVI, 1883, 27e p. 173. 22 "'Major Rogers reports that there is no question about feasibility of good line with easy grades through Kicking Horse Pass although work will be very expensive.'" Van Home, Telegram to Drinkwater, cited by Tupper, Official Report of the  Debates of the House of Commons of the Dominion of Canada, Ottawa, Thenceforth "HoC Debates,"T~April 17, 1882, p. 953. 23 W. Vaughan, op. cit., p. 80. 62 24 Sir Thomas Shaughnessy to R. Douglas, Secretary, Geographic Board, Ottawa, March 23, 1921, Department of Railways and Canals, Railway Branch, Central Registry Files, PAC. RG 43 A 2 (a) 6710 Vol. 223. In the same letter, Shaughnessy averred that "the first Directors and the Executive of the Canadian Pacific considered the route via the Yellowhead Pass too far to the North and involving undesirable length of line." See also, J. L. McDougall, Canadian Pacific: A Brief History, Montreal, McGill University Press, 1968, p. 69; Shaw, op. cit., p. 11; Sprague, op. cit., p. 296. 25 Van Home to Tupper, DSP, Vol. XVI, 1883, 27 1 p. 7. 26 Ibid. 2? "The river has its canons [sic], and in places washes the base of the mountains, so that heavy work and possibly some tunnelling would have been encountered on the longer route." Keefer, op. cit., p. 75. Vaux, op. cit., p. 73, goes so far as to claim that it was "the cost of tunnelling and bridging" alone which persuaded the C.P.R. to seek a direct route across the Selkirks. Published evidence does not support this view. See, Van Home to Tupper, op. cit. 2 8 "'The worst that can happen in case of failure to cross Selkirk is, that the line may be forced round the great bend of the Columbia, which would considerably increase distance..." Van Home, Telegram to Drinkwater, op. cit. 29 Van Home estimated the distance around the Big Bend at 140 miles, and the distance by the direct crossing over Rogers Pass at 63 miles, yielding a saving in distance via the latter of 77 miles. Van Home to Tupper, op. cit., p. 7. Howard Palmer estimated the distance around the Big Bend at 150 miles, and, using the estimate of 63 miles over Rogers Pass, obtained an estimate for the savings via the direct route of 87 miles. Palmer also estimated that the distance from Winnipeg to Kamloops via Edmonton, the Yellowhead Pass and the Albreda Pass was 1,346 miles, against an estimate of 1,224 miles via Calgary, the Kicking Horse Pass, Rogers Pass and Revelstoke, yielding a saving for the southerly route of 122 miles if Rogers Pass were adopted, and of 35 miles if the Big Bend were followed. H. Palmer, "Early Explorations for the Canadian Pacific Railway," Bulletin of the Geographical Soc iety of Philadelphia, Vol. XVI, No. 3, July 1918, p. 78. 3 0 See note (22) . 31 Van Home, Telegram to Drinkwater, op. cit. 32 W. Moberly, "The Introductory Chapter in the History of the Canadian Pacific Railway," Moberly Papers, Vancouver City Archives, (henceforth "VCA,") p. D915. 33 Ibid. Moberly himself proposed for the following year, "A trial survey across the Selkirk Range by the valleys of the Gold 63 River and Gold Creek, to ascertain what length of tunnelling would be required to connect those valleys."' N. Robinson, Blazing the Trail Through the Rockies: The Story of Walter  Moberly, Vancouver, News - Advertiser, printers, 1913, p. 75. 34 '"The crossing of the Selkirk Range is the only thing in doubt, but explorations have progressed sufficiently to justify belief that they can be crossed by use of some long tunnels.'" Van Home, Telegram to Drinkwater, op. cit. 3 5 Ibid. See note (28) . 36 Drinkwater to Tupper, September 15, 1882, DSP, Vol. XVI, 1883, 27, p. 25. 37 Ibid. 38 Stephen to J. H. Pope, Acting Minister of Railways and Canals, September 29, 1882, DSP, Vol. XVI, 1883, 27e, p. 168. 39 Ibid. 40 Drinkwater to Tupper, September 15, 1882, op. cit. 41 Rogers, Engineer, Mountain Division, to Van Home, January 10, 1883, DSP, Vol. XVI, 1883, 27e, p. 171. 42 Van Home to Tupper, April 18, 1883, DSP, Vol. XVI, 1883, 27 1, p. 6. The C.P.R. would use this reasoning to defend an even more drastic upward revision on the western slope of the Rockies, from ninety feet per mile to 116 feet. Ibid., pp. 6-7. 43 Ibid., p. 7. 44 See notes (20) and (21). 45 Stephen, Telegram to Marquis of Lome, September 1882, quoted in Pugsley, op. cit., p. vi. 44 As Howay, and particularly Pugsley, maintain that it was. See F. W. Howay, British Columbia From the Earliest Times to the  Present, Vol. II, Vancouver, S. J. Clarke Co., 1914, p. 424; Pugsley, op. cit., p. 11. 47 See, for example, Drinkwater to Tupper, February 21, 1883, op. cit. 48 Rogers to Van Home, op. cit. 49 Rogers to Van Home, November 20, 1883, DSP, Vol. XVII, 1884, 31f, p. 40. 50 Ibid. 51 "Memorandum of the General Character of the Work, Prepared from the Last Information at Command," Schreiber, February 1, 64 1884, DSP, Vol. XVII, 1884, 31f, p. 43. 52 Ibid. 53 The railhead crossed the Columbia in October 1884, and reached Beavermouth in November. Lamb, op. cit., p. 119. 54 Reed to Van Home, September 9, 1884, DSP, Vol. XVIII, 1885, 25n, p. 5. 55 As late as February 1884, one member of the House of Commons would complain that no cost estimates had been submitted for any of the work west of the summit of the Rockies. HoC Debates, February 18, 1884, p. 359. 56 Ibid., p. 458. 57 Reed to Van Home, op. cit., p. 5. 58 Ibid. 59 $36,927.08 per mile for the entire section, compared with $36,557.38 per mile across the Selkirks. The former average is calculated from an estimate of $10,635,000 for the 288 miles from the summit of the Rocky Mountains to Savona's Ferry. This estimate is contained in "Schreiber's Estimate, Summit of Rocky Mountains to Middle of Eagle Pass," enclosed with, Ross to Van Home, October 7, 1884, "Van Home Letterbooks," Vol. 7, p. 928. The latter average, from the same source, is based on an estimate for the total cost of construction between Miles 1,039 and 1,100, west of Winnipeg, of $2,230,000. Ross presumed that these estimates were "intended to be entirely safe." Ibid., p. 927. The estimates per mile for each section may be calculated from Schreiber, op. cit., as follows:- Mile 963 (summit of Rockies) - 966, $26,250; 967 - 975, $155,555.56; 976 - 1,024, $37,755; 1,025 - 1,038 (the crossing of the Beaver River, the point at which the rail line diverged from the Columbia River), $60,714.28; 1,039 - 1,057 (summit of Rogers Pass at 1,054), $35,789.47; 1,058 - 1,072, $36,666.66; . 1,073 - 1,100 (Revelstoke, the point at which the Columbia River was rejoined), $35,714.24. Ibid. These estimates correspond exactly with those contained in "Progress Estimate No. 56, Central Section, Eastern Division," November 4, 1884, DSP, Vol. XVIII, 1885, 25a, p. 90, except for the estimate of the cost of the section between Miles 967 and 975. In the Dominion Sessional Papers, the estimate for the total cost is given as $400,000, yielding an average cost per mile of $44,444.44. Schreiber's estimate for the total cost of the nine-mile section is quite distinctly stated as $1,400,000, yielding an average cost per mile of $155,555.56. 60 This capital cost includes an estimate of $450,000 for snowsheds on the direct crossing. See note 93. The capital invested on the Big Bend route would have had to have been less 65 than $2,680,000 (i. e. $2,230,000 + $450,000). If the distance around the Big Bend were 140 miles, the cost per mile would have had to have been less than $19,142.86. If the distance were 150 miles, the cost would, have had to have been less than $17,866.67. These rates would have been unprecedented for mountain railway construction .to main-line standards. 61 Stephen to Lome, op. cit. 62 "Official Memorandum Respecting the Position and Prospects of the Canadian Pacific Railway," Sir George Stephen, December 12, 1882, DSP, Vol. XVI, 1883, 27n, p. 5. 63 Ibid. 64 DSP, Vol. XVII, 1884, 31g, pp. 52-53. 65 DSP, Vol. XVII, 1884, 31z, pp. 250-54. 66 Van Home to Rogers, May 23, 1884, "Van Home Letterbooks," Vol. 6, p. 251. 67 0. S. A. Lavallee, op. cit., p. 174. 68 Van Home to C.P.R. Directors, DSP, Vol. XVIII, 1885, 25n, p. 2. 6' See above, p. 44. TO n— a distance which, as everyone familiar with railway management knows, is extremely convenient for the application of a pilot engine." Tupper, HoC Debates, May 4, 1883, p. 960. See also, Drinkwater to Tupper, September 15, 1882, op. cit., and, Rogers to Van Home, January 10, 1883, op. cit. 71 Drinkwater to Tupper, September 15, 1882, op. cit. 72 Stephen to Pope, op. cit. See note (42). 74 Van Home to Tupper, op. cit., p. 8. TS "<rhe heavier gradients, which will in no case exceed those of the Central Pacific Railway, will be confined to the mountain section, and within a space of 150 miles. "It is also to be noted that the entire mountain section is embraced within a distance of less than 550 miles from the Pacific coast, while that of the Central and Union Pacific Railways covers about 1,250 miles and lies at a much greater elevation." "Official Memorandum," by Stephen, op. cit. 76 The phrase is Van Home's, used in describing the adequacy for traffic purposes of the "temporary" 4.5 per cent, gradient on the western slope of the Rockies. Van Home to Minister of 66 Railways and Canals, May 19, 1884, DSP, Vol. XVIII, 1885, 25a, p. 10. 77 Ibid., p. 11. 78 Van Hofne to Tupper, op. cit., p. 7. 75 "Official Memorandum," by Stephen, op. cit., p. 9. 80 Van Home to C.P.R. Directors, op. cit., p. 2. 81 "I do not hesitate to say...that every part of the line, from Montreal to the Pacific, will pay." Ibid., p. 3. 82 Van Home to Tupper, op. cit., p. 8. 83 "I have shown that a large amount of ore or base metal will be shipped from the Kootenay mines over the C.P.R....It will be a valuable trade for that railway, as the transportation will be westwardly, while the bulk of their other freight will be in a contrary direction." G. B. Wright to J. H. Pope, June 11, 1883, British Columbia Board of Trade, Annual Reports, Victoria, 1882-83, p. 31. "With the fast-approaching completion of the Canadian Pacific Railway, whereby direct and speedy transport eastward will be secured, the food-fish trade of this Province must receive a notable impulse. . .A large demand will necessarily arise throughout the line of the railway, where settlement has been established, and in Manitoba; and eastward again of the last named locality, in.Ontario and elsewhere, it is probable that, during the winter season, some of our sea-fishes may prove abundantly attractive, and find a ready and lucrative market." op. cit., 1883-84, pp. 96-97. 84 Van Home to Tupper, op. cit., pp. 7-8. 85 Moberly cited "avalanches of snow and rock" as a principal reason for eschewing direct crossing of the Selkirks. Moberly Papers, op. cit., p. D910. He himself recalled running for his life, clad in snowshoes, to avoid interment in a snow slide. Ibid., p. D916. See also Fleming, op. cit., pp. 264-65. 8 * Rogers to Van Home, January 10, 1883, op. cit. 8 7 See note (42) . See below, pp. 70-72. "Memorandum by Mr. Smellie, Engineer in Chief at Company headquarters, Montreal, dated April 15, 1882," cited by Tupper, HoC Debates, April 17, 1882, p. 954. 90 Reed to Van Home, op. cit., p. 6. 91 Ibid. 67 52 Report of Commission for examination of Union and Central Pacific Railroads, October 30, 1869, DSP, Vol. XIV, 1880-81, 23o, p. 119. 93 Stephen to Minister of Railways and Canals, March 18, 1885, DSP, Vol. XVIII, 1885, 25cc, p. 6. 94 Ross to Van Home, November 23, 1883, DSP, Vol. XVII, 1884, 31f, p. 41. 95 See, for example, DSP, Vol. XVIII, 1885, 25a, pp. 10-16. 96 See, for example, Schreiber to Bradley, November 13, 1884, DSP, Vol. XVIII, 1885, 25a, p. 32. 97 DSP, Vol. XVIII, 1885, 25a, pp. 18-22. 68 CHAPTER 4 REALITIES The C.P.R. Management: "This is the climax of mountain scenery."1 The C.P.R. Customer: "It is not too much to say that the Canadian Pacific passage through the mountains is the greatest sermon ever presented to man on the Divine Majesty. The artist is inspired, the lover of nature satiated."2 The C.P.R. Employee: "In the winter it was snow and frost. In the spring it was snowslides, washouts and every other sort of trouble known to railroading, and in the summer it was fires. Just one continual round of pleasure — if one liked that sort."3 The detailed expectations of the C.P.R. for the Rogers Pass route having been analysed in the previous chapter, the purpose of this chapter is to examine the realities which were encountered in Rogers Pass, that is, the actual conditions of both construction and operation which prevailed along the route. Such an examination is necessary to ascertain the existence of gaps between expectations and realities. Once the existence of any gaps has been established, the specific objectives of remedial measures undertaken by the C.P.R. to close these gaps may be more clearly understood, and the success of those measures may be more readily evaluated. Several compelling narrative accounts have been written of the realities of construction through the Selkirks.4 Evidence of the methods by which railway operations were conducted through Rogers Pass is, however, more fragmented. In order to preserve the analytical character of the present study, and to facilitate the identification of gaps between expectations and realities, 69 the investigation of realities in this chapter will be similar in structure to the investigation of expectations which was undertaken in the final section of the previous chapter. The same five areas of constructional and operational concern will be explored. In the consideration of operating realities, the analysis will be extended to include examination of remedial measures adopted by the C.P.R., for it is recognized that, by their very nature, "operating realities" are not static but dynamic in character. They may change with each traffic movement, each technological or managerial innovation, and it is management's unremitting task to seek to narrow those gaps between operating expectations and operating realities as far as they are able. It is therefore appropriate that Part I of this thesis should conclude in this chapter with some consideration of the extent to which those gaps were narrowed by C.P.R. management in Rogers Pass. a) The Character Of Construction Work Construction work in the Selkirks was dominated by three constraints, those imposed by snowslides, financial pressure and time pressure. The nature of these constraints will be more closely examined below.5 This section will concentrate on the impact which these constraints had upon the alignment which was actually followed. In order to secure cheap and rapid completion of a route over the summit which could be made safe later, as traffic developed, the Construction Manager, James Ross, had intended to 70 undertake "temporary work in the way of building a line that can be thrown further into the hillsides afterwards."6 On the east slope of the Selkirks, this approach could be implemented successfully.7 However, Ross was quickly forced to concede that the avalanche problem had been seriously underestimated,8 and that relocation would be required on the west slope in the interests of safety.' The alignment initially proposed by Rogers had descended the west slope on the north bank of the Illecillewaet River, that is, directly across the south-facing bluffs of Mount Cheops. (See map II.) From observations conducted during the winter of 1884-85, it was discovered that these bluff's, exposed to the sun, "were literally an almost continuous avalanche path."10 The plan to undertake "temporary work (which) could be used to work into the permanent line" proved untenable.11 Ross estimated that "it will take 8,350 feet of shedding and about 1400 feet of tunnelling to operate the line with any safety over these... slides."12 The previous year, in order to avoid the capital cost and delay of building a 1,400-foot tunnel on the west slope of the Rockies, the C.P.R. had obtained the permission of the Federal Government to construct a temporary line with gradients of double the maximum permitted in its contract.13 A year later, confronted with an even more drastic shortage of capital and an even more pressing need to generate revenue from through traffic,14 they lacked resources of both time and money to invest in a 1,400-foot tunnel and 8,350 feet of shedding beneath.snowslides. « 71 • MAP II: LOCATION OF ALTERNATIVE ALIGNMENTS PROPOSED AT CONSTRUCTION TIME ON THE C.P.R. MAIN LINE IN ROGERS PASS. 72 An alternative location was sought on the south bank of the Illecillewaet. However, in order to reach the valley floor and avoid crossing the highly active Ross Peak slide path, a gradient much steeper than the contractual maximum of 2.2 per cent, compensated would have been required.15 Even if the C.P.R. had been prepared to increase its ruling gradient, it is doubtful whether the Federal Government, after having so recently acceded to the controversial request for a temporary line in the Rockies, would have granted permission for a second deviation from the contract in the mountains.16 Instead, Ross developed the line up the valley of Five-Mile Creek, a tributary of the Illecillewaet, .and by inserting an elongated loop into the alignment, contrived to reach the valley floor within the contractual maximum gradient.17 The south bank of the Illecillewaet was less prone to avalanches than the north bank.18 Moreover, the "Loop" itself carried the line into the centre of the Illecillewaet Valley by means of five trestles, which had an aggregate length of 4,108 feet.19 Not only was the new alignment thus clear of slides from both sides of the valley, but construction of the trestles represented a far more rapid and less capital-intensive alternative than tunnelling and snowshedding. Ross estimated that his location would cost "some four to five hundred thousand dollars less to make it a safer line,"20 and that it had "several hundred degrees less curvature upon it" than upon the original.21 However, the sharpest of these curves was 10° 30' at its central angle,22 which was in excess of the contractual maximum 73 of 10°, and the remainder of the Loop was built at that maximum. The Federal Government had urged that curvature be reduced to eight degrees wherever gradients exceeded sixty feet per mile.23 Whilst Van Home had complied on the Kamloops Lake section,24 Ross warned that in the Selkirks the cost of compliance would be "very heavy."25 Construction work on the new alignment was "very heavy" too,26 contrary to the expectation of generally moderate work through the Selkirks. Finally, development of the line had added over three miles to the length of 2.2 per cent, gradient,27 and all of the additional distance opposed eastbound traffic. Thus, the pusher gradient system within the Selkirks was diseguilibrated. Henceforth, the pusher gradient on the west slope would be 24.5 miles long, against 21.5 miles on the east slope.28 The disequilibrium favoured westbound traffic, the direction forecast by the C.P.R. to preponderate. The extent of the imbalance would not seriously increase the operating cost of eastbound movements relative to westbound movements. Nevertheless, the existence of the imbalance between the pusher gradients meant that, if the relative balance of traffic flows did accord with C.P.R. forecasts, the preponderance of the eastbound flow might eventually pose a capacity problem which would be more acute than it would have been had the pusher gradients themselves been balanced. The influence of capital and time constraints, which had dictated the relocation west of the summit, was also manifest in the character of.the bridge- and tunnel-work undertaken through the Selkirks. Bridging, with "good but uncreosoted timber,"25 represented, a rapid, low-capital-cost alternative to filling or 74 diverting streams, although it imposed high subsequent maintenance costs.30 In the 46.1 miles between Beavermouth on the east slope and Albert Canyon oh the west, there were no less than 207 bridges, with an aggregate length of 19,349 feet.31 Five of these were each over one thousand feet long.32 However, perhaps more indicative of the margin at which the C.P.R. were prepared to trade off immediate construction costs against delayed operating costs is the fact that eleven of these bridges were only six feet long, and 118 were sixteen feet long or less.3 3 The unforeseen necessity for tunnelling on the north bank of the Illecillewaet having been avoided by the location of the Loop, tunnelling requirements were largely as expected. No tunnels were necessary on the east slope of the Selkirks.34 On the west slope, where a maximum of 1,200 feet had been projected,35 two tunnels, the Laurie Tunnels,36 aggregating 1,251 feet in length, were eventually constructed between Rogers Pass and Albert Canyon.31 Due to the time and capital constraints, the side-drift method of construction was adopted in order to secure more rapid completion of the first of these,38 and the second may not have been completed until at least a year after the main line opened.3' Although, as will be demonstrated below, construction work proved both more capital-intensive and more time-consuming than anticipated, in practice neither the work which was undertaken nor the alignment which was adopted appear to have differed greatly from expectations. The single exception was the unforeseen necessity to relocate the line over the Loop 75 immediately west of the summit. Ross, at least, would have preferred not to have been obliged to make the trade-off between construction costs and operating costs - in this way.40 Nevertheless, the Loop alignment was less capital-intensive than the original location, and did, therefore, represent a solution to the trade-off decision which was consistent with the construction policy dictated to the C.P.R. by financial circumstances.41 Moreover, the Loop, unlike the "Big Hill" on the west slope of the Rockies, was never perceived as being merely a "temporary" alignment, intended to be improved as soon as the flow of revenue traffic permitted. Rather, it was seen as a satisfactory solution in itself to the trade-off decision, and as permanent a solution as the remainder of the alignment across the Selkirks. Thus, the C.P.R. would claim that, "the general alignment, outside the loop was much improved,"42 and that, "the line as now located is as favourable as any that can be obtained crossing the Selkirks."43 The latter statement can only be accepted in the context of the predominant capital and time constraints which will be examined below. For taken at face value, the statement implies that the trade-offs between construction costs and operating costs would have been handled in exactly the same manner had no such capital and time constraints prevailed. b) The Cost Of Construction The C.P.R.'s entire mountain section, from the Rockies' westward, was constructed under conditions of severe capital rationing. These conditions were already in effect when the 76 railhead reached the Rockies,44 but they were at their most severe during the period of construction across the Selkirks. It was not primarily the capital cost of the actual construction across the Selkirks which was responsible for intensifying these conditions. Three other factors were responsible. These were the capital cost of construction across the Rockies, the inopportune timing of federal loans to assist the C.P.R., and the deliberate reallocation of capital by the C.P.R. from the mountain section to the Lake Superior section in order to accelerate completion of the latter. In March 1884, with the railhead • at the summit of the Rockies, a $22.5 million federal loan had been granted to the C.P.R.4 s By November 1884, however, with the railhead at Beavermouth, the base of the east slope of the Selkirks, the C.P.R. were again "lamentably hard up for money."46 In March 1885, the Company was briefly unable to meet its wage obligations,47 and relief, in the form of a further federal loan, was not secured until July 20th,48 by which time the railhead had almost crossed the Selkirk summit.45 Meanwhile, completion of the Lake Superior section was accorded priority as a means of obtaining further financial assistance.50 Capital savings of some four million dollars were effected on construction work in B.C.,51 and diverted from the mountain section to be "rapidly absorbed on the Lake Superior Section."52 Only after completion of the latter, in May 1885, could the entire capital resources of the C.P.R. be concentrated upon construction of the remaining sections in the Selkirks and the Gold Range. 77 The intensification of capital rationing across the Selkirks fostered the propensity to trade off construction costs against operating costs in a manner which would minimise the immediate requirement for capital. Cuttings were reduced to widths less than those on the sections built directly by the Federal Government,53 against the standards of which Van Home would rail in later years. Ballasting was omitted, and bridges were built entirely of timber, without masonry or iron support.5 4 Despite the conditions of capital rationing, however, and despite the deliberate management policy of minimising immediate construction costs, the estimate of capital requirements for construction across the Selkirks was exceeded. The excess was incurred chiefly upon the west slope. Between the Beaver River and the summit on the east slope, the actual cost corresponded closely with the estimate of November 1884. From the summit to the first crossing of the Illecillewaet on the west slope, however, the November 1884 estimate of $550,000 was exceeded by some $200,000, or over one third. This increase was incurred not so much because of an increase in the capital cost of the work per mile, but rather because three additional route miles had to be inserted into the section at the Loop. Thus, the per-mile cost of $41,666.66 exceeded the per-mile estimate of $36,666.66 by only fourteen per cent. A similar margin of error prevailed on the section between the first crossing of the Illecillewaet and the second crossing of the Columbia, where the length of railway constructed corresponded with the length estimated. Thus, the expected cost, one million dollars for the twenty-78 eight miles, was exceeded by $130,000, or thirteen per cent. The total cost of construction from the Beaver River to the Columbia River, as assessed after completion of the entire transcontinental main line, was $2,560,000. This was $330,000, or fifteen per cent., more than the last estimate submitted before construction across the Selkirks began.55 Certainly, as James Ross admitted, "It has cost more than it should."56 Certainly too, Ross had had "to strain every point to so change the location as to save every dollar."57 Nevertheless, it appeared that this was successfully accomplished. For example, had the location of the Loop not obviated the expenditure of a further $500,000, the actual cost would have exceeded the expectation by some $830,000, or over thirty-seven per cent. C.P.R. management's policy in railway construction across the Selkirks had been overtly to "build the road, in all respects, with the least immediate outlay necessary to insure safety in operation, leaving as much as possible to be done in the future."58 This policy had been dictated by financial expediency. Yet it was the construction policy which was most appropriate for the C.P.R. as it sought to open up a link between east and west across the mountains of B.C.; and the policy was successfully implemented. c) Time Required For Construction Van Home's hope to reach the second crossing of the Columbia by the end of 1884s' was frustrated by the nature of the construction work which was encountered on the west slope of the Rockies. The railhead did not reach Beavermouth, ten miles 79 west of the first crossing of the Columbia, until November 1884 . 6 0 Nevertheless, Van Home's forecasted date for completion, September 1885,61 was actually brought forwards by a month in October 1884,62 perhaps in anticipation of easier work in the Selkirks, or perhaps in an attempt to spur Ross to even greater efforts. Considerable pressure was exerted upon Ross in order to accelerate completion.63 However, work was delayed early in 1885 by avalanches, the first to be experienced directly by Ross and the C.P.R., and as a result of the experience, the Construction Manager counselled Van Home to "anticipate a delay in construction."64 In June 1885, the forecasted date for completion was set back again by two months, to the end of September.65 Heavy summer rainfall further delayed the work, and by late September the railhead had progressed only as far as Albert Canyon, twenty-four miles west of the Selkirk summit, having advanced scarcely fifty miles in six months.66 The "Last Spike" was not driven until November 7th, in Eagle Pass. Even then, however, the line was not opened to through traffic for another eight months. No snowsheds had been provided in either the Selkirks or the Gold Range at the time of construction, and no rail operations were attempted until further research had been undertaken into the incidence of avalanches in the mountains. Only when the slide season was over, and after the damage to the permanent way had been repaired, was the line opened for revenue traffic, in June 1886. This delay in opening the line had been fully anticipated by the C.P.R., and both shareholders and the Federal Government had 80 been forewarned.67 Although the erection of snowsheds to protect the permanent way would continue until 1890, the through line itself had indeed been completed only two months behind the schedule anticipated when the railhead reached the Selkirks. The circumstances causing delay, the snowslides and inclement weather, might possibly have been foreseen. Nevertheless, the delay was trivial in comparison with the saving achieved over the deadline initially contracted. Moreover, the date of actual completion was safely within the limits of the revised contract of 1884. No specific value for "construction-time saved" was adduced in the west.68 However, the time savings were clearly accorded an implicit value, for certain construction decisions, notably the location of the Loop and the building of bridges using timber only, were motivated by the desire to secure time savings as well as by the desireoto secure capital savings. Both types of savings had to be traded off against the increase in subsequent operating costs, and it appears that in the Selkirks, trade-offs intended to secure time savings were successful. In order to save one year of construcion time in the Rockies, the C.P.R. were compelled to locate thirteen miles of explicitly "temporary" track, and to operate over its 4.4 per cent, gradients until a permanent line was built.6' In contrast, no part of the route across the summit of the Selkirks was ever regarded as "temporary" at construction time: when the C.P.R. secured time and capital savings in the Selkirks, they did so without incurring any tacit obligation to undertake subsequent relocation. 81 d) Operating Methods And Traffic Flows Gradients across the Selkirk summit nowhere exceeded the 116-feet-per-mile maximum which had been anticipated by the C.P.R. The 2.2 per cent, compensated ruling gradient commenced on the east slope at Beavermouth, 21.5 miles from the summit, and on the west slope at Albert Canyon, 24.5 miles from the summit. Trains were hauled by a single road locomotive westbound from Field to Beavermouth and eastbound from Revelstoke to Albert Canyon.70 When these trains were too heavy to be hauled over the 2.2 per cent, ruling gradients by the single road locomotive, pusher locomotives were attached in order to avoid the expense of cutting and remarshalling the trains. One pusher locomotive was attached to each ascending train, and provided assistance as far as the yard at Rogers Pass on the summit of the Selkirks.71 Here, while the road locomotive conducted the train on the descent of the opposite slope, the pusher locomotive was detached, turned in the Rogers Pass roundhouse and returned light down the gradient, to be turned again on a "Y" at Beavermouth or Albert Canyon and held to await the next ascending train. The level of the variable cost of hauling traffic over the pusher gradients was determined by the number of pusher locomotives required in the mountain locomotive fleet, and by the number of trips obtained from each pusher locomotive. These latter variables were in turn functions of the volume of traffic requiring transit over the pusher gradients, and the volume of traffic which the locomotives could haul per unit of time, as conditioned by their tonnage ratings and speeds, and by the 82 spacing of meeting and passing sidings along the route. The C.P.R. had expected light traffic for several years after the opening of the transcontinental line, and it appears that they were able to provide sufficient pusher capacity to move the available traffic until at least the turn of the century. Scheduled passenger service was provided by a single passenger train daily in each direction until 1902.72 Passenger trains comprised five or six cars,73 and by 1891, observation cars were being attached through the mountains.74 However, the heaviest of all passenger vehicles, dining cars, were not attached until 1909.75 The speed of passenger trains through the Selkirks was not determined by the tonnage rating of the road locomotive, but was regulated to coincide with a mealbreak at the Glacier Hotel, two miles west of Rogers Pass.76 Pusher locomotives were therefore rarely required on passenger duty. Data limitations inhibit accurate estimation of the volume and direction of freight movements over the Selkirks. However, an impression of low traffic volume and surplus rail capacity may be gleaned. Throughout the summer of 1888, when seasonal freight activity might have been expected to have been at its highest for the year, there were no westbound freight trains scheduled out of Donald at all, and only two per day eastbound.77 Admittedly, most freight trains were run as specials, without scheduling. Yet in 1889, the Superintendent of the Pacific Division observed that "there will be some days without any trains whatsoever over the road, and others, after the arrival of a steamship [i.e. in Vancouver], there will be a constant quick succession of trains for several days 83 together."78 During one such local peak, in a single week of July 1888, the Donald Truth boasted of "seventy car-loads of tea" destined eastbound through the mountains.75 Even this, however, would amount to scarcely one train per day.80 Movements of the other principal traffic from the Orient, silk, may not have been of sufficient volume to generate a demand for more trains: adequate surplus capacity existed on the passenger trains to permit attachment of cars of silk to their rear.81 When avalanches in the Rockies disrupted transcontinental service for three days in December 1888, only one special freight and two tea trains were delayed.82 As late as June 1896, total train movements between Donald and Kamloops, including summer passenger and freight specials as well as mixed trains, amounted to 268, an average of four and a half trains per day in each direction.83 The maximum length of trains over the Selkirks appears to have been at least thirteen cars when the line first opened.8* By 1898, the maximum length of a train with a single pusher locomotive was certainly eighteen cars,85 and this maximum continued until at least 1900.86 Eighteen loaded cars were equivalent to 846 tons, the haulage capacity of a single "Consolidation" locomotive on a 1.6% compensated gradient.87 Over the Selkirks, therefore, where the gradients were 2.2% compensated, pusher locomotives may have been required for any train of greater weight than approximately 420 tons, or about nine loaded cars. It is not known how frequently in practice trains were assisted to the summit. However, evidence suggests that by the 1890's, the operating cost of the pusher service was 84 decreasing, and that this trend continued until the turn of the century. By 1893, the average length of freight and mixed trains on the entire Pacific Division chad risen beyond nine cars, the approximate payload for which pusher service was required, to 10.95 cars. One year later, this had increased to 11.31 cars,86 and in September 1894 it reached 12.13 cars.8' Although these data are for average conditions only, they indicate that payloads per train were continuing to increase beyond the point at which pusher service became necessary, and therefore that the cost of the pusher service on each train was being spread over more revenue traffic.90 Moreover, the C.P.R. does not appear to have experienced any difficulty in providing the requisite pusher capacity. When the line had opened for traffic in 1886, two locomotives had been appointed for pusher duty on either slope of the Selkirks.'1 By August 1898, it had still not been found necessary to increase the size of this pusher fleet.92 Train speeds over the Selkirks were low. The first scheduled passenger trains averaged less than twelve miles per hour for the seventy-nine-mile journey between Revelstoke and Donald.93 By 1902, the eastbound "Imperial Limited," the fastest C.P.R. train through the mountains, still required almost four and a half hours for the crossing, averaging less than eighteen miles per hour.94 The scheduled Third Class "Fast Freight" services averaged just over eleven miles per hour for the journey in 1902.95 It is unfortunate that little is known of the speeds at which the "extra" freight trains negotiated the Selkirk gradients, for it was these services which conveyed the majority of traffic through the mountains. The evidence of 85 passenger-train timetables is unrepresentative of typical train speeds over the Selkirks. These services were rigorously scheduled, and received priority over all other traffic. Their meets were carefully synchronised, and their speeds through the entire mountain region were governed to ensure timely arrival for breakfast, lunch and dinner at the C.P.R. hotels in North Bend, Glacier and Field. This regulation of their speed, and the fact.that several minutes could generally be found for a stop to admire the Illecillewaet River at Albert Canyon,'' suggest that there was little difficulty in pathing the "extra" freight trains among the scheduled services between Beavermouth and Albert Canyon. Van Home had forecast that in the early years, the flow of traffic would be predominantly westbound. Although at first more trains were scheduled to run eastbound than westbound, and although the tea and silk flows which have been so highly visible to subsequent commentators'7 were also always eastbound through the mountains, it appears that Van Home's forecast was accurate for at least several years after the transcontinental facility opened. In July 1888, the Donald Truth lamented that the C.P.R. could not be made to "see that it would be better for it to haul loaded cars east to Winnipeg, rather than empty ones."'8 The Vancouver Board of Trade recorded that in 1889, 38,895 tons of freight arrived in Vancouver by rail from the East, and 21,441 tons were shipped by rail to the East. In 1890, the respective volumes were 50,773 tons and 13,973.5 tons.5' It is unlikely that this westbound predominance was reversed until the "take-off" of the lumber trade between B.C. and the prairies 86 after the turn of the century.100 With the volume of traffic, the length and speed of trains and the frequency of train movements detailed above, it appears that siding capacity on the main line over the Selkirks was quite adequate for the extent of operations which the C.P.R. was required to undertake. By August 1889, there were certainly crossing points for trains at Bear Creek and Glacier, five miles respectively on the east and west slopes from the yard at Rogers Pass.101 Table 1, taken from the earliest available comprehensive list of sidings on the Pacific Division, compiled in 1896, indicates that, despite the nature of the terrain, the C.P.R. contrived to locate passing sidings at regular intervals on the main line, and that each of the sidings could easily accommodate the longest trains operated through the mountains. The frequency of sidings and the ratio of siding length to length of running line on the C.P.R. in 1896 compared favourably with those on the Canadian Northern main line when the latter opened through the Yellowhead Pass in 1915.102 87 TABLE 1 SIDING ACCOMMODATION IN THE SELKIRK MOUNTAINS, C. 1896 Section Length  Main Line ^Miles!-Length  Side Track (Feetl Car-lengths  Of Storage (36' Per Car*) Revelstoke East Twin Butte West Twin Butte East Albert Canyon West Albert Canyon East Illecillewaet West Ross Peak West Ross Peak East Glac ier Rogers Pass Bear Creek Six-Mile Creek Beaver West Beaver East Donald 5 1/2 5 5 5 1/2 5 5 5 5 5 5 5 1/2 5 6 6 800 22 2,645 2,785 1,150 12,016** 3,850 2,000 1,000 1,000 7,171 17,012 73 77 31 333 106 55 27 27 199 472 * Standard length of C.P.R. box-car. See, "Hunting-Merritt Lumber Company versus Canadian Pacific and British Columbia Electric Ry. Companies, 20, Canadian Railway Cases, 181 at 184 ** Includes 9,500 feet of summer track Source:- Abbott to Shaughnessy, September 15, 1896, PIC, CPCA 88 The^C-P'R«'s expectations for operating conditions and traffic flows through the Selkirks proved accurate once the line opened. The volume of traffic was low, and a small fleet of pusher locomotives was sufficient to provide assistance over the summit. Despite slow speeds and the necessity for returning pushers light against ascending trains, traffic levels were sufficiently low, and siding accommodation sufficiently spaced, to permit the pathing of trains through the Rogers Pass corridor with little apparent difficulty. Although it is not known how intensively the available pusher capacity was utilised, it seems that paths for additional trains could have been found had increasing traffic made them necessary and had the pusher fleet been enlarged accordingly. It is therefore concluded that there was surplus line capacity on the C.P.R. route over the Selkirk Mountains. It appears from this analysis that that surplus capacity endured until at least the turn of the century. e) Snowslide Protection In adopting the location through Rogers Pass, the C.P.R. had expected to encounter snowslides for some ten to twelve miles, and had expected to solve the problem by building snowsheds. During construction of the transcontinental line, however, in the C.P.R.'s first winter in the Selkirks, Ross admitted candidly that he had underestimated the danger from the slides,103 although he reassured Van Home that conditions in the Selkirks that winter were exceptional.104 The C.P.R.'s realisation of the magnitude of the threat may have been further postponed due to the fact that in the following winter, that of 89 1885-86, when they sent observers into the Selkirks to monitor slide paths, the slides were "certainly less in bulk" than they had been the previous winter.105 In the winter of 1886-87, conditions were far worse than they had been in either of the previous two years,10' and even these conditions may have been surpassed in adversity by those of 1887-88.107 While Ross acknowledged the gap between the expectations of snowslides and the realities, he was confident that the gap could be closed with modest immediate cost,108 and that, "With some additional expense over estimate every point can be made perfectly secure for operation."109 He proposed to minimise the investment in snowshedding, the immediate cost, by maintaining an increased section force who would dig out the slides "as you would an ordinary drift."110 However, not only was the gap between expectations and realities underestimated, but so was the cost of closing that gap. Even after the experience of the first winter's observations, Van Home informed the Ministry of Railways and Canals that, "A comparatively small expenditure will be required to make the line entirely safe."111 When, as a result of these observations, the General Superintendent of the Pacific Division submitted estimates for the cost of snowshedding, Van Home was forced to admit that, "Your estimate of cost...is far beyond any previous estimate and far beyond our expectations, and for financial reasons it is somewhat appalling."112 The General Manager, consistent with his expectation that the line would be lightly used for several years, intended to reduce the immediate cost of closing the gap, and to postpone further investment in 90 snowslide protection, until traffic volumes had increased. He was prepared to countenance risk in order to make the trade-off in this way. He informed his General Superintendent that, We can't afford to cover every place where a slide may occur. If we provide now for such as have occurred within record years we will probably be justified in taking some chance of interruption by the others.113 The fact that this was the C.P.R.'s trade-off policy in providing protection from snowslides makes even more remarkable the magnitude of the sums of capital which were invested in the construction of snowsheds during the years immediately following the completion of the through line. After Ross's reports during the winter of track-laying, the C.P.R. had estimated that $450,000 would suffice for snowsheds in the mountains.114 During 1886, the first summer after observation of the avalanches, the C.P.R. spent $1,477,510 on snowsheds in the Pacific Division.115 The following year, having estimated that a further $504,565 would be required,116 the C.P.R. was in fact forced to invest $691,062,117 and in 1888 the Company disbursed another $136,401.116 Perhaps because of the magnitude of the costs, or perhaps because the C.P.R. might have argued that investment in snowsheds was a social welfare measure, it seems that by April 1888, the Company had applied for, or was at least hopeful of, a Federal Government grant for snowshedding work.11' However, there is no record of such a grant having been approved. Two factors may explain why the C.P.R. was unable to contain investment in snowshed construction within target levels, and why its trade-off policy broke down in the provision of avalanche protection. The first factor was the extent of snowshed construction. After the first winter's observations, 91 thirty-five sheds were constructed.120 In 1887, a further 15,388 feet of shedding were proposed, 8,568 feet of which were to be located in the Selkirks.121 By June 1888, a total of 31,764 feet of sheds had been built on the entire Pacific Division, of which 30,403 feet, or 5 3/4 miles, were situated between Beavermouth and Albert Canyon in some forty-three separate sheds.122 The second factor was the quality of construction which was required if the sheds were to function satisfactorily. In order to withstand the force of avalanche material weighing from 25 -45 lbs. per cubic foot,123 the protective structures, sheds and glance-cribs, had to be far more heavily built than those on the American transcontinental railways. Snowsheds on the Central Pacific had cost an average of $10.25 per lineal foot in the 1860's.124 In 1918, the renewal cost of sheds on the Southern Pacific would be between twelve and fourteen dollars per lineal foot.125 The average cost of the C.P.R. sheds and cribs by 1888 was around seventy dollars per lineal foot. The need to upkeep this quantity of shedding imposed severe maintenance costs upon the C.P.R., which will be considered in more detail in the following chapter. This extent of shedding also created operating problems of its own, however. Each shed had to be patrolled constantly by section men, in winter on account of avalanche damage, and in summer on account of fire damage, to which the sheds were peculiarly susceptible.12' The longer sheds were equipped with hydrants every four hundred feet.127 Due to the steep main-line gradients, handcars could rarely be used in fire-fighting,128 and eventually a fleet of locomotives had to be adapted for this purpose.125 During 92 routine operations, the accumulation of smoke rendered the brakesmen's duties hazardous,130 and in winter the rails were prone to icing and the permanent way deteriorated.131 As the C.P.R. acquired experience of the slide paths, they undertook bridge improvements and piecemeal diversions as alternative means of preventing the avalanches from obstructing the main line through the mountains. The trestle bridge at Snow Bank was swept away in February 1886 , 1 3 2 and again in January 1887.133 New bridges were installed here and at Cut Bank, enabling avalanches to pass beneath the railway tracks.134 When the bridges were again removed by slides, Cut Bank in 1900 and Snow Bank in 1904,135 the respective ravines were filled, and the main line diverted in each case. Diversion at Williamson's Creek, some two hundred yards east of Cut Bank, as part of the Cut Bank project, cost $1,242.57 in 1900.136 Two years previously, a diversion at I llec i llewaet had cost $9 , 429'. 91.13 7 These cost data suggest that diversion was adopted where it clearly represented a less costly means of avalanche defence than the snowsheds. However, these investments in bridge improvements and diversions must not be accounted entirely as costs of snowslide protection. Rather, they may have been the result of a greater availability of capital once transcontinental operations had commenced, and they may have formed part of the policy, clearly envisaged at the time of construction, of investing in the upgrading of the line once traffic had begun to flow.138 These investments in improvements, which had the effect of mitigating the avalanche problem, were therefore entirely consistent with the manner in which the 93 C.P.R. had sought to handle the trade-off between construction costs and operating costs from the very outset of work in the Selkirks. Where diversions were too costly, and where the snowsheds failed, the line was cleared by snowploughs and the efforts of the section gangs. The first winter of operations proved that the conventional wing ploughs were "entirely insufficient and almost unworkable" in the Selkirks,13' and the C.P.R. was compelled to invest in the more expensive rotary snowploughs. The first was delivered to the Selkirks and tested in November 1888.140 It was, however, required to perform duty not only in the Selkirks, but in Eagle Pass too.141 The single plough appears to have been inadequate for the burden of these snowclearing duties. Nevertheless, C.P.R. management clearly accorded greater priority to the clearing of the Lake Superior section, and refused requests for a second rotary plough for the mountains "until it is known what difficulties are likely to be encountered this winter on the North Shore."142 It was not until February 1890 that a rotary was transferred from the Lake Superior section to the Mountain Division, by which time, "Want of a second rotary ha(d) seriously delayed operations in clearing snow slides."143 The rotaries alleviated the difficulty of finding convenient dumping grounds for the cleared snow,144 but the early models were nevertheless useless for clearing avalanches containing timber and rocks, and these had still to be cleared manually. Assembling an adequate labour force in the mountains could be a task in itself.145 Certainly, the C.P.R. underestimated both the demand for 94 snowslide protection in the Selkirks and the cost of meeting that demand. Nevertheless, within five years they had implemented a comprehensive avalanche defence system intended for the protection of their main line. This- system had preventitive components, diversion of the line, snowsheds and patrols, and curative components, snowploughs and section gangs. Moreover, the system was successful. The C.P.R. itself claimed that the sheds "answered their purpose admirably," and that during the winter of 1889-90, the first in which the defensive system was fully operational, the C.P.R. "was the only one of the transcontinental lines that enjoyed immunity from blockades."146 Available evidence reinforces these claims. Daily records of the arrival time of the C.P.R. passenger service in Vancouver reveal that from November 1888 to January 1890, only two trains were cancelled "on account of obstructions in the mountains."147 Two others were more than twelve hours late in reaching their destination.148 From January to December 1891, three trains were more than twelve hours late, all in late April, and in December one train was cancelled, although it is not known whether this cancellation was due specifically to avalanche problems in the Selkirks.145 The Minister of Railways and Canals assured the House of Commons in May 1888 that, ...the means adopted by the (C.P.R.) Company for dealing with the avalanches of snow in the Rocky Mountains [sic] were found to be absolutely perfect, the snow-shedding, which is upon a scale that would astonish hon. gentlemen if they were to see in the solidity of construction, allowing these avalanches to come down from the Rocky Mountains and the Selkirks and elsewhere to pass over them without the slightest difficulty or without the slightest disturbance."150 Even Walter Moberly, who spent a lifetime deploring the adoption 95 of the Kicking Horse Pass and Rogers Pass in preference to the Howse Pass and the Big Bend, was forced to concede that, The most direct line is unquestionably the one taken by the C.P.R. along the Illecillewaet river. Its disadvantages are the heavy grades and liability to snow and land slides. Substantial snowsheds are overcoming one of these difficulties.151, 152 Moreover, it should be remembered that slide problems at least equal in severity to those in the Selkirks were encountered elsewhere through the mountains. In January 1887, Abbott reported that, "the difficulties with the snow have occurred where we least expected them, viz. at Eagle Pass,"153 and after the 1887 slide season, he accompanied his estimate for snowshed requirements with a recommendation that, If the Company decide upon building a portion only of these sheds, I would suggest that those in Eagle Pass should be first provided, so as to confine the trouble to the line between Revelstoke and Donald, and that the worst of the slides in the Selkirks should then be provided for, according to the amount that may be appropriated for this purpose.154 A list of delays to passenger trains submitted in March 1889 indicates that rock slides between Vancouver and Kamloops were responsible for most of the lost time, whilst in the Selkirks, "the glance cribs, fences, sheds &c. have all stood the test and are doing the work, for which they were intended, admirably."155 In November 1892, Abbott would report forty-eight mudslides in a single day on the Thompson and Cascade sections,15' and when, in 1894, the year of the C.P.R.'s darkest fortunes, their main line through B.C. was closed for forty-one days, it was not avalanches in the Selkirks which were responsible, but flooding in the Fraser.15 7 96 The preceding analysis of the areas of constructional and operational concern to the C.P.R. reveals a close correspondence between expectations and realities. The gap between expectation and reality in the nature of construction work west of Rogers Pass did not translate into a significant gap in terms of either cost or time. Operations were conducted, and traffic conveyed, much as expected. The only serious gaps between expectations and realities emerged over the incidence of avalanches and the cost of measures to protect against them. The analysis reveals that in undertaking these remedial measures, the C.P.R. endeavoured to handle the trade-off decision in providing snowslide protection exactly as they had handled it in construction of the line as a whole: that is, they sought to minimise capital investment at the outset, and to undertake measures of improvement as operations over the line developed. To this extent, the extremely high capital cost of snowslide protection should be regarded as an indication, not that the route over the Selkirks was fundamentally unsafe, but simply that the C.P.R.'s trade-off policy was, in the specific context of avalanche defences, inappropriate. The policy broke down. There was no effective solution to the snowslide problem which was not capital intensive. Thus, It was deemed best to carry out these works in the most durable and substantial manner, in order that the safety of the line might be placed beyond doubt.158 Even though the C.P.R. had been forced to abandon their intended trade-off policy, they had made, as it were, the minimum concession. Having once grasped the magnitude of the snowslide problem, and the cost of solving it, the Company was 97 faced with four alternatives. They could abandon the Rogers Pass route entirely, and build around the Big Bend; they could follow the Rogers Pass alignment, but undertake tunnelling beneath the snowslides, presumably to the extent of the 2 1/2 miles which they had been prepared to sanction in order to secure a direct crossing; they could persist with the alignment as built, but abandon operations during the winter months;15' or they could persist with the alignment as built, and invest substantially in snowsheds for its protection. To have either abandoned the Rogers Pass route entirely or to have undertaken tunnelling would have imposed demands upon the capital resources of the Company, and indeed upon the capital resources of the country, which quite simply could not have been met in 1885, nor in the years immediately afterwards. There is no evidence that the Company ever considered these alternatives. There is no evidence that the Federal Government ever asked them to consider the alternatives. The Federal Government, indeed, would not even contribute to the cost of the snowsheds. To have closed the line in winter would have meant foregoing the revenue from all traffic which might have traversed the line during the months of closure. It would also have entailed investment in the spring in order to repair the damage inflicted by the avalanches of the winter. After the closure of 1885-86, the damage to the unprotected line had not been repaired until the following August.160 Not only might the investment in repairs be high, therefore, but interruption to traffic as a result of leaving the line unprotected might not 98 have been confined to the winter months alone. The fact that the C.P.R. rejected this alternative suggests that they believed that the direct cost of repairing the line after closure, combined with the opportunity cost of interrupting the flow of traffic, together outweighed the cost of constructing and maintaining avalanche defences in order to keep the line open throughout the winter months. Therefore, the Company proceeded to invest heavily in snowsheds. A gap between expectations and realities had existed. Remedial measures were taken. These measures were successful. The gap was narrowed. Indeed, insofar as the line was never disrupted for more than a month once the avalanche defence system was implemented, the gap must be regarded as having been almost entirely closed, and the trade-off decision must be regarded as having been largely successful. There is one final implication of the result of the C.P.R.'s selection from among the four alternatives outlined above. The decision to attempt to keep open the line over the summit of the Selkirks on a perennial basis implicated the C.P.R. in a high fixed investment in snowsheds. The investment was particularly high, and particularly fixed, in comparison with other investments undertaken by the C.P.R. in construction through the Selkirks. Even discounted back to 1885, at four per cent.,1'1 the amount invested in snowsheds by 1888 still represented $2,180,869, almost as much again as the investment in the main line itself between the Beaver and Columbia rivers, and as much as the C.P.R. had expected to pay for construction through the Selkirks. This fixed investment had been undertaken, 99 and annual maintenance charges would accrue, regardless of the volume of traffic which actually took advantage of the line's being open during those winter months on behalf of which the costs were incurred. The implication must be that, in the years immediately following completion of the transcontinental link, the C.P.R. was in a decreasing-cost situation in the Selkirk Mountains. Specifically, the more traffic that could be carried through the Selkirks, the wider the fixed cost of snowshed maintenance could be spread, and the lower the total cost of each individual traffic movement would be. Moreover, given the low initial volume of traffic travelling over the line, additional traffic could be handled without incurring congestion costs. Therefore, the reduction of total costs effected by spreading the snowshed costs over an increased volume of traffic would not be offset by the addition of more trains. Thus, although the C.P.R. had handled trade-off decisions in a manner intended to ensure low capital costs, it commenced operations in a situation where it could absorb additional traffic without incurring increased operating costs. 100 FOOTNOTES 1 Canadian Pacific Railway Company, "The Canadian Pacific; the new highway to the East, across the mountains, prairies and rivers of Canada," Montreal, 1888, p. 25. 2 J. B. Ker, "The Progress of Vancouver," in, Vancouver Board of Trade, Annual Reports, Vancouver, 1892, p. 23. 3 Alex Forrest, quoted in E. E. Pugsley, "Pioneers of the Steel Trail. Four: Fighting the Snow Menace," Maclean's Magazine, August 15, 1930, p. 16. 4 Perhaps the most compelling is that of 0. S. A. Lavallee, Van  Home' s Road, op. cit., pp. 194-214. 5 See sections (e), (b) and (c) respectively. 6 Ross to Van Home, March 4, 1885, quoted in Lavallee, op. cit., p. 196. I Five years after the line opened, one traveller would record ascending the east slope "along a track cut in the side of the mountain..." Francis Mollison Black, "Down the Selkirks on the Cowcatcher: A Story of Rogers Pass," MSS, VCA, July 1891, p. 1. 8 "I find that the snowslides on the Selkirks are much more serious than I anticipated, and I think are quite beyond your ideas of their magnitude and of the danger to the line." Ross to Van Home, February 19, 1885, quoted in Lavallee, op. cit., p. 194. ' "I can see quite plainly that the present location of the line will not be safe — more particularly so on the west slope where the slides this season already aggregate more than two miles in width." Ibid. 10 Proceedings of the Canadian Society of Civil Engineers, op. cit., p. 27. II Ross to Van Home, February 19, 1885, quoted in Lavallee, op. cit., p. 196. ...to get any kind of line we have to go in for heavy work which would in no way serve our purpose in throwing the snow..." Ibid. 12 Ibid. 13 DSP, Vol. XVIII, 1885, 25a, pp. 10-14. 14 The C.P.R. was on the verge of bankruptcy by March 1885, and relief in the form of federal aid was not forthcoming until July. Lamb, op. cit., pp. 128-132, H. A. Innis, A History Of The  Canadian Pacific Railway, Toronto: McClelland and Stewart, 1923, pp. 125-6. 101 15 Haldane estimated "the natural slope of the line" on the south bank as one in 17 1/2, equivalent to an uncompensated gradient of 5.7 per cent., or approximately 300 feet per mile. J. W. C. Haldane, 3,800 Miles Across Canada, London: Simpkin, Marshall, Hamilton, Kent & Co. Ltd., 1900, p. 216. 16 Vaux implies that the Federal Government did in fact refuse to sanction the steeper gradient. Vaux, op. cit., p. 84. 17 DSP, Vol. XIX, 1886, 35a, p. 11. 18 "On my way west, I noticed on the other side of the Ille-Cille-Wait [sic] that there were no large slides or any marks of very dangerous ones..." Ross to Van Home, March 4, 1885, quoted in Lavallee, op. cit., p. 199. 15 The length of the trestles was as follows: First Crossing, Five-Mile Creek, 331 feet; Second Crossing, Five-mile Creek, 1,006 feet; First Crossing, Illecillewaet, 601 feet; Second Crossing, Illecillewaet, 1,061 feet; Third Crossing, Illecillewaet, 1,109 feet. "List of Bridges," Kilpatrick MSS, Vancouver, 1893. The map in "Snow War, A guide to the history of  Rogers Pass, Glacier National Park," Ottawa: Dept. of Indian and Northern Affairs, Parks Canada, 1978, p. 8, erroneously labels the First Crossing of the Illecillewaet east of Glacier House. According to the "List of Bridges," this bridge was called "Glacier Creek," and was 211 feet long. The "List of Bridges" quite clearly records the First, Second and Third Crossings of the Illecillewaet as west of the two crossings of Five-Mile Creek. 20 Ross to Van Home, March 25, 1885, quoted in Lavallee, op. cit. , p. 199. 21 Ibid. 22 Ross to Van Home, June 18, 1885, ibid., p. 205. 23 Schreiber to Bradley, February 8, 1883, Department of Railways and Canals, Railway Branch, Central Registry Files, PAC. RG 43 A 2 (a) 6710 Vol. 223. 24 DSP, Vol. XVIII, 1885, 25a, p. 32. 2 s "The Government should be asked to accept ten degrees as the minimum to Station 1200 West of the Summit, otherwise the increased cost will be very heavy." Ross to Van Home, February 19, 1885, Presidents' Inward Correspondence, Canadian Pacific Corporate Archives, Montreal, (henceforth 'PIC CPCA,'). 26 Schreiber to Bradley, July 6, 1885, DSP, Vol. XIX, 1886, 35a, p. 7. 27 20,006 feet, according to Lavallee, op. cit., p. 199. 28 C.P.R. Co., "Pacific Division Time Table No. 1, to take 102 effect One 0' Clock Saturday, July 3rd, 1886." Calgary Tribune Print. 25 Lavallee, op. cit., note to Plate 299, p. 183. 30 See below, p. 92; pp. 243-4. 31 "List of Bridges," op. cit. 32 Three were trestles in the Loop. The others were Mountain Creek Bridge, 1,086 feet, and the Fifth Crossing of the Illecillewaet, 1,091 feet. Ibid. 33 Ibid. 34 Rogers to Van Home, November 20, 1883, DSP, Vol. XVII, 1884, 31f, p. 39. 35 Ibid. 36 Lavallee, op. cit., footnote to p. 205. 37 "Tunnels on Pacific Division," n..d., Kilpatrick, Add Mss 323, PABC. 38 Ross to Van Home, June 18, 1885, quoted in Lavallee, op. cit., p. 205. 39 Ross employed the side-drift method on the 565-foot tunnel. Ibid. (This tunnel is shown on the list of "Tunnels on the Pacific Division, op. cit., as 564 feet long.) He reported that he was "running a temporary line around" the other. Lavallee, op. cit., p. 205. On November 10, 1886, the General Superintendent of the Pacific Division reported a runaway incident, with three fatalities and six injuries, "on the middle of the temporary steep grade west of the summit at Rogers Pass." Abbott to Van Home, November 10, 1886, PIC, CPCA. This "temporary steep grade" may have been Ross's "temporary line" around the second Laurie Tunnel. 40 "For my own part I regret being obliged to submit this line but there are so many objectionable features on the present location and the more you examine them, the less you like them..." Ross to Van Home, March 25, 1885, quoted in Lavallee, op. cit., p. 199. 41 See below, pp. 75-78. 42 Schreiber to Bradley, October 10, 1885, DSP, Vol. XIX, 1886, 35a, p. 11. 43 Schreiber to Bradley, July 6, 1885, ibid., p. 7. 44 It appears that the rate of construction across the prairies was faster than optimal. Lamb argues that austerity dictated changes in the proposed alignment west from Calgary as early as 103 mid-1883. Lamb, op. cit., p. 104. 45 DSP, Vol. XVII, 1884, 31z, pp. 250-254. 46 Van Home to John Ross, November 29, 1884, "Van Home Letterbooks," Vol. 8, p. 888. 47 Lavallee, op. cit., pp. 199-204. 48 Lamb, op. cit., pp. 129-132; McDougall, op. cit., pp. 61-63. 49 The railhead crossed the Selkirk summit on August 17, 1885. Lavallee, op. cit., p. 209. 50 "...there is still much lack of faith on the part of the Government and in financial circles of our ability to finish our work within the amount of the Government loan and we will be utterly unable to get any financial relief from outside until the last spike is driven in the Lake Superior section, and the lie is given to all the slanderous reports that have been circulating." Van Home to John Ross, op. cit., pp. 892-3. 5 1 5 2 Blake, HoC Debates, June 20, 1885, p. 2749. Van Home to John Ross, October 19, 1884, quoted in Lamb, op. cit., p. 127. See also Stephen to Minister of Railways and Canals, March 18, 1885, DSP, Vol. XVIII, 1885, 25cc, p. 3. 53 Van Home to H. J. Cambie, July 14, 1885, "Letterbooks," op. cit., Vol. 6, pp. 919-920. 54 Van Home to Schreiber, December 1, 1884, "Letterbooks," op. cit., Vol. 8, pp. 951-2. 55 "Central Section, Western Division, Progress Estimate No. 87, November 28, 1885," DSP, Vol. XIX, 1886, 35a, p. 152. 56 Ross to Van Home, April 16, 1885, quoted in Lavallee, op. cit., p. 204. 57 Ibid. 58 Van Home to H. J. Cambie, op. cit., p. 920. 5 9 See above, p. 54. 60 Lamb, op. cit., p. 119. 61 See above, p. 54. 62 "We ought to be able to complete all the grading by the first of July and to connect the track by the first of August." Van Home to John Ross, October 20, 1884, "Letterbooks," op. cit., Vol. 8, p. 229. 63 For example, before Ross had even completed work on the west 104 slope of the Rockies, and with only two months remaining of 1884, Van Home wrote to him, "I presume upon reaching the Columbia you will be able to lay track, not alone to the mouth of the Beaver, but up as far as the first of the high trestles. It is important that every inch possible should be made this year." Van Home to Ross, ibid., p. 222. 64 Ross to Van Home, February 19, 1885, quoted in Lavallee, op. cit., p. 194. 45 Canadian Pacific Railway Company: Report of the Directors of the C.P.R. Co. submitted at the adjourned Annual General Meeting of the Shareholders, June 13, 1885. Canadian Pacific Railway Company, Annual Reports, Montreal, 1885, p. 18. 66 Lavallee, op. cit., p. 209. 67 As early as March 18, 1885, Stephen had forecast to the Minister of Railways and Canals "the opening of the through line in the spring of 1886." Stephen to Minister of Railways and Canals, March 18, 1885, DSP, Vol. XVIII, 1885, 25cc, p. 1. C.P.R. shareholders were assured in January 1885 that, "by the early spring of next year the through line from Montreal to the Pacific Ocean...will be finished and in perfect condition..." C.P.R. Co., Annual Report, op. cit., 1885, p. 18. In October 1885, with only thirty-six miles of track remaining to be laid, Schreiber informed the Ministry, "I do not think it is the Company's intention to operate (the road) through the mountains this season; in fact I should not consider it wise to attempt to do so until the road is thoroughly completed, which will scarcely be before spring." Schreiber to Bradley, October 10, 1885, DSP, Vol. XIX, 1886, 35a, p. 11. 68 Unlike on the Lake Superior section, where "Mr. Stephen estimates the financial advantage of connecting the track in March instead of May at $500,000." Van Home to John Ross, op. cit., pp. 894-5. Van Home to Minister of Railways and Canals, May 19, 1885, DSP, Vol. XVIII, 1885, 25a, pp. 10-11. 70 Lavallee is incorrect in stating that pusher locomotives were provided from Revelstoke to Rogers Pass. Lavallee, "Rogers' Pass: Railway to Roadway," Canadian Rai1, Canadian Railroad Historical Association, No. 137, October 1962, p. 155. See Marpole to Shaughnessy, February 15, 1898. PIC, CPCA. 71 The pusher locomotive was generally attached at the rear until 1907. This arrangement of the motive power ensured even distribution of the pulling and buffing forces throughout the train. It may also have permitted attachment of the pushers "on the fly." See T. H. Crump, "The Big Hill and the Mountain Section," October 21, 1940, reprinted in Canadian Rail, No. 275, December 1974, p. 356. The change to double-heading in 1907 was ordained by G. T. Bury as General Manager of Western Lines, ostensibly in the interests of safety and passenger comfort. The 105 rearrangement of motive power per se appears to have had a negligible impact upon the economics of the pusher operation. Sir George Bury, "The Making of a Railway Man II. From Superintendent To Vice-President," Maclean's Magazine, January 15, 1926, p. 14. 72 Province, April 21, 1902, p. 1. 73 Crump, op. cit., p. 356. 74 F. M. Black, op. cit., p. 1. 75 "Diary, 1909," Kilpatrick MSS, Vancouver, December 9, 1909. 76 C.P.R. Co., Time Table, July 3rd. 1886, op. cit. 77 Donald Truth, July 7, 1888, p. 8, and November 3, 1888, p. 3. 78 Abbott to Van Home, January 5, 1889, PIC, CPCA. 79 Donald Truth, July 14, 1888, p. 5. 80 Six trains per week, assuming thirteen cars per train, and four trains per week, assuming eighteen cars per train. See below, notes (84) and (85). 81 "(The steamship Aberdeen) had several carloads of silk, one of which went through Donald on yesterday's express, another going through today." Donald Truth, July 28, 1888, p. 5. 82 Whyte, Telegrams to Van Home, December 18, 19 and 21, 1888, PIC, CPCA. 83 Tait to Shaughnessy, July 29, 1896, PIC, CPCA. 84 Donald Truth, October 6, 1888, p. 1. 85 Marpole to Shaughnessy, February 15, 1898, PIC, CPCA. 86 Tye to Shaughnessy, April 2, 1900, PIC, CPCA. 87 Ibid. 88 Memorandum by Thomas Tait to Shaughnessy, October 24, 1894, PIC, CPCA. 89 Shaughnessy to all General Superintendents, October 23, 1896, "Letterbooks," op. cit. 90 The data do not necessarily imply the availability of surplus capacity, or underutilised "push," on each train. Although the data which is based on averages indicate that pusher locomotives were required for the "average" train, and that the marginal payload for which the pusher was required was only, for example, 1.95 cars or 92 tons in 1893, it is possible that the payloads of the trains to which pushers were attached may have been far 106 greater than 10.95 cars, while the payloads of those trains which were conducted by a single locomotive may have been far less than nine cars. Under these conditions, the "push" would be more fully utilised where provided, even though the data for average payload appears to indicate otherwise. 91 Report of H. Abbott, quoted in Lavallee, Van Home's Road, op. cit., p. 244. 92 "Appropriations, Year 1898," Kilpatrick MSS, p. 55. The total haulage capacity of the fleet may have been increased during this period if stronger locomotives were introduced into mountain service. Unfortunately, this hypothesis cannot be tested until the appearance of a comprehensive work on C.P.R. motive power, currently in preparation by 0. S. A. Lavallee. However, since the numerical strength of the pusher fleet certainly did not increase during this period, it is likely that the increment in total pusher capacity obtained from the introduction of stronger locomotives would not have been dramatic. 93 C.P.R. Co., "Timetable No. 1, 1886," op. cit. 9* Canadian Pacific Railway Company, "Timetable Number 1, Taking Effect at 24.01 O'clock, Sunday, June 15th, 1902," Montreal, n .p. 95 Ibid. 96 C.P.R. Co., "The Canadian Pacific: the new highway to the East," op. cit., p. 27. 97 See, for example, Lavallee, op. cit., p. 280; N. R. Hacking, Hi story of the Port of Vancouver, Vancouver, n. d., n. p. 98 Donald Truth, July 28, 1888, p. 4. 99 Vancouver Board of Trade, Annual Reports, op. cit., 1889, p. 31; 1890, p. 24. 10 0 See chapter 6. 101 Abbott to Shaughnessy, August 2, 1889, PIC, CPCA. 102 The average spacing on the C.P.R. between Albert Canyon and Beaver East was one siding every 5.8 miles in 1896, and the ratio, "length of siding: length of main line" was 1: 8.17. Between Lucerne and Blue River, west of the Yellowhead Pass on the Canadian Northern main line, the average spacing was one siding every 8.5 miles, and the ratio, "length of siding: length of main line" was 1: 15.33. Sessional Papers of the Province of  British Columbia, Victoria, 7 Geo. 5, 1917, Report of Department of Railways, p. D12. This does not of course indicate that the C.P.R. had "more capacity" than the C.N. The above comparisons take no account of such crucial variables as average train weight and average train speed. 107 103 Ross to Van Home, February 19, 1885, quoted in Lavallee, op. cit., p. 194. 104 "From all reports the snow is exceptionally deep this season." Ross to Van Home, March 4, 1885, PIC, CPCA. 105 Cunningham, op. cit., p. 21. 106 "Journal of Observations in camp three miles east of Selkirk Summit, Winter 1885-86, kept by Granville C. Cunningham, Engineer-in-Charge (and by J. S. Vindin after April 18th); Observations at Cascade Camp, Winter 1886-87, kept by J. E. Griffith." PABC, pp. 46-57. On February 28, 1887, six C.P.R. employees were killed in an avalanche off Mount Carroll, ibid., p. 54. The following day, the entire distance from snowshed No. 5 to Rogers Pass, some five miles, was "one continuous slide." ibid., p. 55. The line was closed from February 26 to March 23. Ibid., p. 56. 107 Marpole, Telegram to Van Home, December 17, 1887, PIC, CPCA. 108 "In [sic] the east slopes slides occur in two places, but very little shedding will be necessary as the increased section force will dig out any of them quickly, the snow keeping soft will do no damage. On the west slopes slides mostly come down in gulches, so it will be necessary to throw the line more into the hill side so as to pass the snow over the track." Ross, Telegram to Van Home, March 4, 1885, PIC, CPCA. 105 Ibid. 110 Ross to Van Home, March 4, 1885, quoted in Lavallee, op. cit., p. 196. 111 Van Home to Bradley, May 6, 1886, Letterbooks, op. cit. 1X2 Van Home to Abbott, July 4, 1886, ibid. 113 Ibid. 114 See chapter 3, note (93). 115 C.P.R. Co., Annual Reports, 1886, p. 36. 116 "Statement of Proposed Snowshed work," enclosed in Abbott to Van Home, April 15, 1887, PIC, CPCA. Of this amount, $386,515 were earmarked for the section between Donald and Revelstoke. 117 C.P.R. Co., Annual Reports, 1887, p. 25. 118 C.P.R. Co., Annual Reports, 1888, p. 23. In 1889, a further $3,975.95 was invested, and in 1890, $159.25, bringing the total expenditure on snowshed construction on the Pacific Division in the first five years of the line's history to $2,309,108.69. 108 119 Abbott submitted a calculation of the amount of money necessary for snowshedding in the event of the Company's obtaining a grant from the Dominion Government. Abbott to Van Home, April 17, 1888, PIC, CPCA. Unfortunately, the calculation has not been preserved, so there is no way of knowing how the Company proposed to allocate the grant. 120 Keefer, op. cit., p. 68. 121 "Statement of Proposed Snowshed work," op. cit. 122 Keefer, op. cit., Plate V. 123 Engineering News, Vol. XIX, January 21, 1888, p. 38. 124 See chapter 3, note (92). 125 Engineering News-Record, Vol. LXXX, January 3, 1918, p. 45. 126 In the Selkirk Mountains, eleven watchmen were allocated exclusively to snowshed patrols: two each to Sheds 1-6, 7-11 and 16-20, and one each to Sheds 12-15, 21-26, 27-31, 35-38 and 39-42. "Timekeeper's Force Return, Donald - Revelstoke, w/e July 13, 1889." PIC, CPCA. 127 Engineering News, Vol. XIX, January 21, 1888, p. 39. 128 Abbott to Van Home, July 31, 1889, PIC, CPCA. 125 Keefer, op. cit., p. 70. 130 Ibid., p. 69. 131 Abbott to Van Home, January 5, 1889, PIC, CPCA. 132 "Journal of Observations," op. cit., p. 27. 133 Ibid., p. 49. 134 Abbott, Telegram to Van Home, January 28, 1888; Abbott to Van Home, January 29, 1890, PIC, CPCA. 135 Canadian Pacific Railway Company, "Old Bridge Record and Section Maps, Mountain Subdivision," Mount Revelstoke and Glacier National Parks, File No. 1758, p. 14. 136 Ibid. 137 C.P.R. Co., Annual Reports, 1898, p. 19. 138 "I feel sure that in a number of places particularly on the section from the Selkirk Summit eastwards five or six miles and from the Summit westwards towards Glacier Creek — the track will have to be thrown far into the face of the slope before it can be fully protected and these possible changes should be kept in view in building sheds." Van Home to Abbott, July 4, 1886, 109 op. cit. 139 Abbott to Van Home, April 15, 1887, op. cit. 140 Donald Truth, November 24, 1888, p. 1. 141 Abbott to Shaughnessy, November 29, 1889, PIC, CPCA. 142 Ibid. . Marpole, Telegram to Shaughnessy, February 13, 1890, PIC, 14 3 CPCA 14 4 Abbott to 0. W. Petri, April 24, 1889, PIC, CPCA. 145 During February 1890, Abbott reported that, "in order to get the necessary force we had to call upon tie-makers, bridge gangs, Siwashes and every man we could find, as men were extremely scarce at that time on this Division." Abbott to Shaughnessy, May 22, 1890, PIC, CPCA. 14' C.P.R. Co., Annual Report, 1889, p. 13. 147 Vancouver Board of Trade, Annual Reports, 1890, p. 21. 148 Ibid., pp. 22-23; 1889, p. 32. X4» Ibid., 1892, pp. 45-46. 150 Tupper, HoC Debates, May 11, 1888, p. 1337. It should be noted that there were no snowsheds, and very few avalanches, on the C.P.R. main line through the Rocky Mountains. 151 Letter from Walter Moberly to the Editor of the "Winnipeg Call," August 24, 1888. Reproduced in the -Donald Truth, September 1, 1888, p. 2. 152 See also, Engineering News, Vol. XXII, December 14, 1889, p. 570. 153 Abbott to Van Home, January 11, 1887, PIC, CPCA. 154 Abbott to Van Home, April 15, 1887, op. cit. 155 Abbott to Van Home, March 2, 1889, PIC, CPCA. 156 Abbott to Van Home, November 29, 1892, PIC, CPCA. 157 C.P.R. Co., Annual Report, 1894, p. 10. 158 C.P.R. Co., Annual Report, 1886, p. 11. 159 There is some evidence that this alternative was considered. Perhaps after the C.P.R.'s experience of its first winter in the Selkirks, the Calgary Daily Herald may have considered that the Company had no other choice: "The town [i. e. Rogers Pass] is 110 built right in the track of the avalanches and after November will be subject to the disturbance of these mountain horrors. At the beginning of December all the inhabitants will move out in a body and Roger's [sic] Pass will be desolate until another summer spreads her mantle on the scene, etc." Calgary Daily  Herald, August 6, 1886. Van Home must have considered discontinuing passenger services during the winter of 1886, either in the immediate interests of passenger safety, or because he was reluctant in the coming months to undertake the full cost of making the line safe for passenger travel, or perhaps simply because of lack of patronage of the passemger service. On December 9, 1886, he informed Abbott, "We have decided on continuing the daily through passenger service for the Winter." Van Home to Abbott, December 9, 1886, Letterbooks, op. cit. 160 Engineering News, Vol. XV, May '8, 1886, p. 303. 161 This was the rate of discount adopted by the C.P.R. in the evaluation of alternative tunnelling projects through the Selkirk Mountains in 1912. See Chapters 7 and 8. It is unlikely that the discount rate would have been significantly less in 1885 than it was in 1912. Ill PART TWO  THE BIG BORE The C.P.R. had made its decision. It had evaluated alternative routes through the mountains, and had opted for a short, direct crossing. In the Selkirks, this entailed construction and operation over Rogers Pass, with its steep gradients and exposure to snowslides. The decision to secure a direct route had been taken in 1881 and fully implemented by 1885. It would commit the C.P.R. to surface operations in Rogers Pass for the next thirty years. What were the consequences for the C.P.R. of being committed to this alignment? One of the principal consequences was that the C.P.R. became engaged in a protracted and costly battle to protect its traffic against avalanches. This consequence has attracted the most scrutiny from previous historians of the C.P.R.'s operations in Rogers Pass. Many of these historians assert that the C.P.R. lost the battle. The 1910 avalanche disaster, in which 62 C.P.R. employees were buried alive, is regarded as a turning point in C.P.R. management's perception of the viability of the surface route through Rogers Pass. The C.P.R.'s decision, taken in 1913, to abandon the surface route and construct the Connaught Tunnel beneath the summit of the Selkirks, is regarded as an acknowledgement of defeat, a strategic withdrawal in reaction to the intractability of the avalanche hazard. How tenable is an explanation of the decision to construct the Connaught Tunnel which addresses only the snowslide 112 problems? For another principal consequence of the C.P.R.'s commitment to a surface route through Rogers Pass was the necessity to haul all trains over 46 miles of 2.2% gradients and severe curvature. The variable cost of routine operations was therefore high. Moreover, the steep gradients, the single-track configuration of the main line, and the scarcity of locations suitable for sidings, all imposed constraints upon the capacity of the facility to absorb increases in traffic. The Connaught Tunnel was double-tracked, and secured a large increment in main-line capacity. It was accompanied by gradient revisions which reduced the variable cost of operations and enhanced the increment in line capacity provided by the tunnel. Analysis reveals that traffic forecasts generated by the C.P.R. in 1913 identified an urgent requirement for increased capacity through Rogers Pass, and that investment in the Connaught Tunnel was undertaken with a view towards expected future operating requirements, and not just in.response to past avalanche experiences. The second part of this thesis examines the question of why the surface route through Rogers Pass was abandoned in favour of the Connaught Tunnel. The answer is sought by an analysis of operating conditions at the summit of the Selkirks throughout the thirty years of surface railroading. This section begins with a re-examination of the nature of the avalanche hazard which has been accorded so much attention by previous railway historians. An attempt is made to establish the actual extent and cost of snowslide problems on the surface route. Then, traffic developments through the B.C. mountains are examined, 113 with particular emphasis upon previous investments undertaken by the C.P.R. to improve operating conditions in Rogers Pass, and upon the implications of traffic growth and traffic forecasts for the future adequacy of the surface route. When the C.P.R. deemed that its surface alignment was no longer appropriate for its operating requirements, it considered several alternative alignments, and ultimately decided to construct the Connaught Tunnel. The C.P.R.'s evaluation of these alternative alignments is described, and the reasons for the selection of the preferred alternative are discussed. Finally, the conclusions of the thesis are presented. 114 CHAPTER 5 AVALANCHE PROBLEMS The purpose of this chapter is to analyse the role played by avalanche problems in the decision of the C.P.R. to abandon the surface alignment through Rogers Pass. Previous historians of C.P.R. operations in the Selkirks have maintained that the role played by snow problems was crucial, and they have made little attempt to look beyond this aspect of operating conditions for an explanation of the decision to construct the Connaught Tunnel. This chapter will attempt to determine whether or not the snow problem in Rogers Pass was indeed sufficiently severe to justify abandonment of the original route over the Selkirks, and whether or not it was indeed the desire to avoid the danger and expense of avalanches which spurred the C.P.R. to invest in an alignment underground. The analysis is divided into two parts. The first part concentrates exclusively upon the role of the 1910 avalanche disaster in motivating the decision to relocate the main line. This concentration is justified because certain authorities attribute the decision entirely to the influence of that particular disaster.1 In the second part of the analysis, the focus is widened to include consideration of the snowslide problem in general, the extent of the problem and the impact which it had upon the investment decisions taken by the C.P.R. in Rogers Pass. 115 5.1 The 1910 Disaster On the evening of March 4, 1910, a C.P.R. snow-clearing crew was working at the south end of Shed 17, one mile west of Rogers Pass station at the summit of the Selkirks. The crew was removing a slide which had descended during the afternoon from Mount Cheops, to the west of the main line. Half an hour before midnight, the crew was struck by a much larger avalanche descending from Mount Avalanche, east of the main line. Sixty-two C.P.R. employees were killed,2 of whom thirty-two were "Japs and Hindoos."3 There were several alarming aspects of the disaster, besides the enormity of the death-toll. The snow-clearing crew had -been working on a two-mile portion of the main line which had been relocated less than three years before. The relocation, motivated by the desire to increase yard accommodation rather than by any necessity for avoiding snowslides over the original location,4 had been undertaken in the belief that the new route was quite safe from snowslides. Indeed, the C.P.R. had undertaken additional investment in widening the cuttings on the diversion,5 in order to secure greater protection.' The incident might have been regarded as an indication that no surface alignment through the Pass could escape the avalanche danger, or that no expansion of capacity could be secured without increasing the vulnerability of the operation to disruption by avalanches. Moreover, whilst no members of the public had suffered injury in the incident, the westbound passenger train No. 97 had been less than ten miles away when the fatal avalanche had struck,7 and had fortunately been running slightly 116 late, having been delayed by a smaller snowslide east of the Selkirk summit.8 The train would be imprisoned in the mountains for two and a half days until the major avalanche could be cleared.' Less than a week before, over eighty passengers on the Great Northern's Spokane Express had been killed at Wellington, Washington, when an avalanche had swept the train into a 150-foot gorge. 10 If the 1910 disaster is to be linked directly with the decision, taken more than three years later, to abandon Rogers Pass, then proponents of the direct linkage must believe that the incident precipitated a change in G.P.R. management's perception of the viability of the route through the Selkirk Mountains. For when the disaster occurred, the C.P.R. had been operating over the surface alignment for some twenty-four years, during which time they had made no serious attempt to seek an alternative route to that through Rogers Pass. In order to establish the extent of any linkage, this analysis will begin by considering the nature of the 1910 disaster itself. Then, the possibility will be investigated that the disaster provoked pressure upon the C.P.R. to undertake investments in improving the safety of its surface alignment. Finally, the actual response of the C.P.R. to the disaster will be examined, for evidence that the incident had indeed prompted a reappraisal of the viability of the surface route. The 1910 avalanche disaster was in every sense a "freak." The slide followed a path down which there had been no previous record of avalanches.11.At an inquest into the deaths, it was stated that, "There was timber in the path of this slide which 117 in some places was fifty years•old."12 The slide was also of exceptional magnitude. An employee of twenty years' seniority testified that, The old track for considerable distance several hundred feet west of 17 Shed was covered by the slide, as well as the new track. In my experience this has not occurred before.13 Finally, the weather in the days preceding the slide was exceptional, even for the Selkirks. "There had been a snowfall of 88 inches in nine days previous to the slide."14 There is no evidence to suggest that the disaster provoked any pressure upon the C.P.R. either to undertake improvements to the existing route, or to abandon the route entirely and invest in an alternative. Certainly, the first coroner's jury which was empanelled to investigate the disaster was dismissed on March 12 after failing to reach a verdict.15 However, the controversy appears to have centred upon the questions of whether or not the C.P.R. actually compelled its snow-clearing crews to work at nights, and whether or not the failure to post look-outs at the site of snow-clearing operations constituted an act of negligence on the part of the Company. At a second inquest, convened on March 14, it was reaffirmed that, "It is not compulsory for men to work at nights."1' Those who did were paid time-and-a-half.17 It was moreover agreed that look-outs would be "not much use at nights" in any event.18 The findings of the inquests placed no pressure upon C.P.R. management to make a major policy decision. The jury at the second inquest returned a verdict of "Accidental Death." It expressed no condemnation of the C.P.R. Neither did it recommend drastic changes in methods of operating through the Pass, nor 118 expensive investments in improving traffic conditions. Rather, it recommended simply that, "... the Canadian Pacific Railway withdraw their workmen from service, at all slides in future during stormy nights."1' There was nothing new in this: the General Superintendent of the Pacific Division himself had issued instructions to this effect as early as 1888 after a C.P.R. work-train had been struck by an avalanche in the Selkirks.2 0 Neither was any pressure from other institutions exerted upon the C.P.R. to undertake investments in improving its surface alignment. The Federal and Provincial Governments made no allusion to the incident,21 and the Ministry of Railways and Canals merely noted, without comment, that the deaths were responsible for inflating the annual accident figure for 1910 to an unusually high level.22 The Labour Gazette reported the incident matter-of-factly in its account of "Industrial Accidents."23 Whilst a motion to investigate industrial safety was carried in the House of Commons within a year of the 1910 disaster, and whilst the frequent incidence of injuries to railwaymen did spark the debate, neither the 1910 disaster in particular nor the character of the C.P.R.'s operations through Rogers Pass in general aroused comment from the protagonists.24 Neither was the press critical of the C.P.R. Indeed, the Company emerged with credit, the Revelstoke Mail-Herald carrying a glowing account of the conduct of the C.P.R. officials during the incident,25 and the Calgary Daily Herald affirming that the C.P.R.'s snow-clearing organization "has been equal to the occasion."2' Nor did the press recommend investment in the 119 pursuit of greater safety. Only the Revelstoke Mail-Herald suggested that the C.P.R. seek an alternative route and accelerate completion of the Arrowhead and Kootenay line.27 The C.P.R.'s own reaction to the disaster sets the incident in its appropriate context as part of the ongoing battle with the snow. The inquests established that there was no question of the C.P.R.'s having omitted to make investments which would have reduced the risk to life and traffic. It had made such investments in the past, and would continue to do so after 1910, without the necessity for major diversions or tunnels: The Company have built sheds wherever they thought it necessary, and have built several sheds over which a slide has never passed... If a shed were thought necessary, expense would not stand in the way.28 The C.P.R., of course, refused to acknowledge any responsibility for the incident, although granting compensation to the relatives of the victims.25 Nevertheless, it appears that the Company was not dissatisfied with its handling of the snow problem. President Shaughnessy wrote privately in the aftermath of the disaster, While it would appear that the danger is not passed by any means, and there is still occasion for much apprehension and anxiety, the record up to the present time is most excellent, marred only by the sad catastrophe, that no human agency could prevent or control, in which so many poor workmen lost their lives.3 0 The C.P.R.'s investment response to the disaster confirms the view that the incident did not provoke a change in investment policy on the Selkirk route. The Company had not built a snowshed over the new alignment because it had not considered it necessary: no slide had ever passed over Shed 17, 120 and it was believed that there was sufficient flat land to the east of Shed 17 "to stop any ordinary slide," even considering that, "A slide had never been known on the [east] side for many years."31 It does not appear that the realignment had in any way entailed increased vulnerability to avalanches as the price of increased capacity. The Resident Engineer at Revelstoke affirmed that, A new track 300 or 400 feet to the north would have been reached by this slide. A shed over this portion of new track would not have withstood this slide.32 Management's investment response to the 1910 slide was the same as it had been to previous slides. "The first thing we will have to do is to build a snow shed at Rogers Pass on the new line," the Vice-President of Western Lines, Sir George Bury, had written to the Chief Engineer on March 15, 1910.33 The shed was erected during 1910, at a cost of $48,275.97,34 and the next year some $700 was invested in a shed over the Rogers Pass turntable.35 Two new rotary snowploughs were ordered, and the existing fleet modified.36 A piecemeal diversion was undertaken at Bear Creek.37 The old alignment through Rogers Pass which the C.P.R. had intended to abandon in 1907, and which it had re connected to the main line two days after the 1910 disaster in order to pass the beleaguered train No. 97,38 was retained "as emergency track in cases of blockades on the Diversion by snow."39 However, there was never any question that the disaster would induce the Company to abandon its new alignment through the Pass.40 The 1910 disaster may have provoked discussion of the possibility of driving a tunnel beneath Rogers Pass, for in 121 April 1910, the Revelstoke Mail-Herald reported that, ...it is stated a tunnel would soon save its cost in the maintenance and construction of snowsheds, besides avoiding the danger of slides in the pass. A factor in the problem is that the time is at hand when all snowsheds and the extensive cribwork connected with some of them would have to be wholly renewed in any case.41 However, there is no evidence to suggest that the C.P.R. regarded such a project as an appropriate alternative to its longstanding policy of avalanche defence. Moreover, analysis of the economics of such a project reveals that in fact the savings in snowshed construction and maintenance would not alone be sufficient to justify investment in a tunnel on the scale which would be required in order to preclude the necessity for shedding through Rogers Pass.42 It is undeniable that the C.P.R. did examine the feasibility of an alternative route through the Selkirks during the summer of 1910, undertaking a thorough survey of the Big Bend. Walter Moberly thought that the time for his favoured route was at hand, and believed that, "The Rogers Pass accident may make [the C.P.R.] change their minds."43 However, the surveys through the Big Bend may have had far grander motives than simply the desire to avoid a repetition of the 1910 disaster. The Victoria Daily Times reported in July 1910, That the Canadian Pacific Railway is in earnest in its scheme to open up the Big Bend by railway transportation and build a connecting line between Revelstoke and the Grand Trunk Pacific at Tete Jaune Cache is evident from the fact that a party of locating engineers numbering sixteen have arrived from the east.44 Moreover, it was chiefly for developmental reasons, and not because the alternatives offered a safer passage through the mountains, that the Revelstoke Mail-Herald welcomed the prospect 122 of rails through the Big Bend,45 just as it had welcomed a report of the C.P.R.'s intention to complete the Arrowhead and Kootenay line.46 It does not appear, therefore, that this survey of an alternative route through the Selkirks was directly linked with the 1910 disaster in Rogers Pass. Regardless of the normative issue, of whether or not the 1910 avalanche disaster should have precipitated a change in the C.P.R.'s perception of the viability of the surface alignment through Rogers Pass, the positive conclusion of this analysis is that the incident in fact heralded no turning point. It was a serious incident in terms of the number of casualties involved, b.ut the slide itself blocked the main line for only two and a half days, a brief interval when compared with the disruptions of previous years, and when compared with the disruption which would follow later in the spring of 1910. The incident brought no condemnation of the C.P.R., and no pressure upon them to undertake investment in improving safety, either upon its existing alignment or by means of an alternative route. It is unlikely that a single incident of this magnitude would stimulate a change in investment policy as drastic as that which would be involved in abandonment of the surface alignment after a quarter-century of rail operations; and in fact no such change occurred. After the 1910 disaster, as before, the C.P.R.'s investment policy towards the avalanche problem continued to be essentially reactive in character. Snowsheds were repaired and extended, piecemeal diversions undertaken, and snowploughs engaged to clear the line between. 123 5.2 The Snow Problem In General Having decisively rejected the hypothesis that it was the 1910 avalanche disaster which induced the C.P.R. to abandon the surface alignment through Rogers Pass, it is necessary still to determine the extent to which the snow problem in general prompted the abandonment decision. Several authorities maintain that it was the apparent worsening of avalanche difficulties during the early years of the 20th Century which persuaded the C.P.R. to undertake investment in a tunnel beneath the summit of the Selkirks.47 In assessing the accuracy of this interpretation, it is useful to distinguish between two separate aspects of the snow problem before attempting to determine whether or not the avalanche difficulties were in fact worsening. These two aspects are the direct cost of maintaining the avalanche defence system, and the indirect cost of disruptions to traffic consequent upon slides blocking the main line. a) The Direct Costs Of Maintaining The Avalanche Defence System Authorities have tended to concentrate upon the direct costs of maintaining the avalanche defence system as the major stimulus to investment in a tunnel. However, quantitative data appertaining to these direct costs do not support this view. Detailed cost data are available for the later years of the surface operation, the crucial years, according to certain authorities, during which the C.P.R. was persuaded "that the savings in snowshed maintenance alone would tip the scales in favour of a five-mile, double-tracked tunnel."48 124 It appears that as the C.P.R. acquired experience of the avalanche problems, and as the Company undertook piecemeal relocations of the line in order to reduce its exposure to snowslides, it was able to reduce the length of snowshedding which had to be provided and maintained. When first constructed across the Selkirks, the line had been equipped with some 30,403 feet of snowsheds,*' and by August 1898 the length of shedding between Beavermouth and Albert Canyon, Sheds 1 to 43A inclusive, had been increased to 31,558 feet.50 It was envisaged in 1898, however, that only some 30,866 feet of this shedding would be renewed, in a programme extending until 1904.51 By October 1904, the last year for which complete data are available, the length of shedding in the Selkirks had been reduced to 29,639 feet,52 a saving of 1,919 feet since 1898. This reduction in the length of snowshedding may have represented a cost-saving of between $6,865 and $8,077 per year.53 Further reductions in the amount of snowshedding on the surface alignment may not have been possible. When the C.P.R. undertook diversion of the main line through Rogers Pass in 1907, it saved another 2,224 feet of sheds,54 which may have afforded annual cost-savings of between $7,956 and $9,360.55 However, after the 1910 disaster, the Company was forced to rebuild Shed 17. With the rebuilding of this single shed, some three thousand feet long, the entire reduction in shedding obtained in 1907 was offset, and the new alignment actually required more avalanche protection than the old. The cost of maintaining and renewing the snowsheds may have escalated rapidly in the last years before the surface route was 125 abandoned. In 1910, the total cost of maintenance and renewals to snowsheds through the Selkirks was $68,481.94.s6 In 1911, the total cost leaped by 75%, or $50,932.05, to $119,413.99.57 Much of this leap may be explained by the rebuilding of Shed 17, which cost $48,275.97. In turn, however, the reconstruction of Shed 17 might have ensured that maintenance costs would have remained at a high level had the surface route continued in operation: for nine months of 1912 the total maintenance cost was $114,878.66. 5 8 Even if this escalation in maintenance costs was entirely due to the rebuilding of Shed 17, which was in turn a consequence of the 1910 disaster, and even if maintenance costs were expected to remain at these high levels for perpetuity, the magnitude of the annual maintenance costs would still not alone have justified investment in a tunnel. In 1912, when the C.P.R. evaluated various tunnelling projects which were intended to supersede the surface alignment, it estimated that 23,760 feet of snowshedding would be rendered obsolete by a tunnel.5' Savings in the maintenance and renewal costs of this shedding were estimated at between $85,000 and $100,000.60 In order to obtain these savings, the C.P.R. would not have been justified in investing more than between $2,125,000 and $2,500,000.61 The lowest estimate for the cost of a tunnel was $5,495,000. 6 2 Even when it became clear that the main line through the Selkirks would have to be doubled,63 the magnitude of the savings which could be derived from avoiding the cost of doubling the existing sheds and maintaining these enlarged sheds would still not alone have justified investment in a tunnel. The 126 C.P.R. estimated that the cost of doubling 23,760 feet of wooden sheds would be $475,200,64 and that the maintenance cost of the enlarged sheds would be $125,000 per year. This increase in the maintenance cost of a doubled shed, from between $85,000 and $100,000 to $125,000, suggests that the C.P.R. believed in the existence of economies of scale in the provision of snowsheds. In order to avoid the capital and maintenance costs of doubling the sheds, the Company would still only have been justified in investing some $3,600,200 . ' 5 Had the C.P.R. merely desired to avoid the costs of maintaining the snowsheds, it could have rebuilt the sheds in reinforced concrete. When this alternative was considered in 1912, however, the estimated cost of double-track sheds in reinforced concrete was $3,801,600.*' The expense of maintaining the wooden sheds was not sufficiently great to warrant this investment: the potential net benefit of such a project, $3,600,200, did not outweigh the cost. The cost of maintaining and renewing snowsheds was the largest single component of the total direct cost of maintaining the avalanche defence system. There were other components, for which quantitative data are not available. The cost of certain of these components, for example the cost of snowshed patrols and the cost of section-gangs clearing line blockages, may have been subsumed within the maintenance and renewal costs discussed above. There is little evidence to suggest that these costs were either significant or escalating. The system of patrols does not appear to have been fundamentally modified throughout the entire thirty years of surface operations. In 1912, patrols were still 127 detailed to Sheds 1-6, 7-11, 12-14, 16-20, 21-27, 28-31 and 35-37,67 much as they had been when the patrol system had been instituted by Van Home." The hours of patrol duty were longer during the summer months than during the winter months, perhaps reflecting the fact that fire was perceived to be a greater enemy of the sheds than avalanches.69 At its most expensive, manning of these patrols could cost up to $620 per month.70 However, even if this rate represented the average for the year, the direct cost of providing the patrol would still only have been $7,440. If this cost were not subsumed within the cost of snowshed maintenance and renewal, it would still amount to less than ten per cent, of the total cost of maintenance and renewal. The cost of other components, for example the acquisition of snowploughs and the diversion of the main line, should not be allocated entirely to the direct cost of avalanche protection. The snowplough fleet, comprising two wingploughs and two rotaries at Revelstoke and one of each at Rogers Pass in February 1904,71 had been augmented by two more rotaries at the end of 1910. 72 The new rotaries boasted significant technological advances over their predecessors. They were, therefore, more than simply reinforcements for the fleet: they represented an investment in modernisation too. Their duties were not confined to the Selkirks, but extended to the Eagle range and the Rockies also. Moreover, they performed not only avalanche clearance but routine snow removal in this region of high winter precipitation. They were acquired, not because the snowslides themselves were increasing in magnitude or frequency, but because, "The increasing traffic makes it most necessary 128 that interruptions be at least cut down to the minimum..."73 The investments were thus motivated, at least in part, by an increase in the volume of traffic over the line. A similar motivation also dictated certain relocations of the main line and bridge improvements which, as has been noted,14 often afforded other operating advantages besides merely the avoidance of snowslides. The role of traffic increases in motivating investments in the mountains will be investigated more thoroughly in the next chapter. This analysis of the direct costs of maintaining the avalanche defence system indicates that the anticipated savings in direct costs were not a major stimulus to the investment in the Connaught Tunnel. The quantifiable costs of maintaining the system certainly appear to have risen in the years immediately preceding the decision to abandon the surface alignment. However, there is no evidence to suggest that the C.P.R. expected these costs to continue to rise, and in 1912, the level of maintenance costs was still not sufficiently high to warrant investment in a tunnel. If the decision to abandon Rogers Pass was motivated by the escalation of the direct costs of maintaining the avalanche defence system after 1910, . then construction of the Connaught Tunnel represented a very expensive solution to the problem. The conclusion that construction of the Connaught Tunnel was motivated by the desire to avoid the direct cost of avalanche protection would only be justified if the sum of the expected savings in the quantified and non-quantified costs outweighed the expected cost of a tunnel beneath Rogers Pass. 129 The maximum quantified saving was $3,600,200. In order to tip the balance in favour of a tunnel, the non-quantified savings, in both direct costs of avalanche defence and indirect costs of traffic disruption, would have had to have exceeded $1,894,800, or over half as much again.75 Given that rotary snowploughs would still have been required in order to remove the routine snowfall, and given the fact that improvements to the permanent way afforded other operating advantages which became increasingly valuable as traffic volumes increased through the Selkirks, it is unlikely that the non-quantified savings in direct costs alone did exceed this level. The conclusion of this analysis, therefore, is that investment in the Connaught Tunnel was not motivated by the direct cost of maintaining the avalanche defence system. It remains to be proven in the next section that the non-quantified savings in indirect costs of traffic disruption would not tip the balance in favour of a tunnel either. b) The Indirect Cost Of Disruptions To Traffic Previous historians of. the C.P.R.'s operations in Rogers Pass have rarely accorded explicit credence to the view that investment in a tunnel beneath Rogers Pass was motivated by the desire to save the indirect cost of interruptions to traffic flows consequent upon snowslides blocking the main line. Nevertheless, the argument has intuitive appeal. This section investigates the possibility that the costs of disruption to traffic provoked construction of the Connaught Tunnel. The nature of the disruption costs is explained, and the incidence 130 of disruption is examined. The efficacy of diversionary arrangements is assessed, and an attempt is made to determine whether the extent of disruption to traffic was increasing in the final years of surface operations through Rogers Pass, as traffic volumes increased. The section ends with an analysis of the importance of disruption costs in the financial evaluation of the Connaught Tunnel. (i) The Nature Of Disruption Costs If traffic flows through Rogers Pass were increasing in the early years of the 20th Century — and, as the following chapter demonstrates, they most certainly were, and at a dramatic rate -- then the indirect cost of line blockages must also have increased. This indirect cost may have had several components. It would certainly have included the cost of diverting traffic via alternative routes, and it would certainly have included the opportunity cost of actually having to forego traffic because of the line blockage. Moreover, if traffic levels were sufficiently high, the indirect cost may also have included a congestion cost, as backlogs of traffic which accumulated during the period of the facility's closure would have been moved under congested conditions once the line could be reopened. Each of these indirect costs would increase as traffic volumes increased. If it is to be argued that the indirect cost of avalanches motivated abandonment of Rogers Pass, it must be proven that the savings to be derived from the avoidance of these indirect costs, either in isolation or in conjunction with savings in the direct cost of maintaining the avalanche defence system, 131 outweighed the anticipated cost of investing in a tunnel. Quantification of these indirect costs would be a difficult accounting problem under any circumstances., but it is rendered particularly exacting in this instance by severe data constraints. There is a paucity of evidence surrounding even the general nature and extent of traffic disruptions consequent upon avalanches, and cost data are virtually non-existent. This analysis, therefore, does not explicitly quantify the indirect cost of disruptions to traffic resulting from snowslides. However, it is at least possible to specify the relationship between disruption costs and investment in avalanche defence. The optimal level of investment in avalanche defence is determined by two variables: the level of avalanche activity disrupting the line, and the level of traffic requiring transit over the line. An increase in avalanche activity in Rogers Pass, or an increase in traffic during the avalanche season, would both increase the probability of delays to traffic, and could both be expected, therefore, to call forth increased investment in avalanche defence, until a new equilibrium between disruption costs and protection costs was reached. Once the optimal level of investment was attained, however, the marginal cost of securing an additional "degree" of protection would be greater than the economic benefits which could be anticipated from the incremental investment. As has been recounted above, the C.P.R. initially provided some 30,000 feet of snowsheds on the main line across the principal avalanche paths. This length of snowshedding was not significantly extended throughout the thirty years of surface operations, presumably because the 132 marginal cost of extension was not justified by the marginal benefit of the additional degree of protection. Moreover, there was a marked discontinuity in the avalanche-defence investment-function, at the point where further investment in snowsheds was abjured in favour of the Connaught Tunnel. The analysis in section (a) of this chapter determined that, at least in the very last years before the decision was taken to abandon the surface operation, the C.P.R. did increase its investment in avalanche defence. It may therefore be assumed that the costs of traffic disruption also increased during this period. The increase in disruption costs may have been sufficient to warrant the increased investment in snowsheds. However, the analysis in the remainder of this thesis reveals that the increase in disruption costs was unlikely to have been sufficient to warrant investment in the Connaught Tunnel. Indeed., it seems likely that avalanche defence exhibited decreasing-cost characteristics throughout the thirty years of surface operations in Rogers Pass. The decreasing-cost nature of avalanche defence is readily explained. If there are only two trains per day over a route, a snowslide may block the line and be cleared again before either of the trains is disrupted. In this case, the provision of a snowshed at the site of the slide averts no disruption, whilst the entire cost of providing the snowshed must be recouped from the revenues of those two trains alone. If, however, there are twenty trains per day over the route, it is unlikely that a snowslide can be cleared before some disruption to traffic occurs. Investment in a snowshed 133 which can withstand the force of an avalanche therefore entirely averts this disruption, whilst the cost of providing the snowshed is spread over all of the twenty trains which travel the route, thus decreasing the total cost of each traffic movement. At the commencement of surface operations through Rogers Pass, the C.P.R. provided some 30,000 feet of snowsheds. Yet in the winter of 1888, only four trains per day were scheduled to cross the Selkirks.76 In the winter of 1912-13, the C.P.R. was still providing some 30,000 feet of snowsheds, yet there may have been as many as fourteen trains daily over the line throughout the avalanche season.7 7 A greater volume of traffic was thus benefitting directly from avalanche defence, whilst contributing greater revenue towards offsetting the costs of the defence system. Between 1910 and 1912, total annual traffic through Rogers Pass virtually doubled.78 Even if expenditure on avalanche defence doubled in the same period — and it is known only that a discrete increase of 75% occurred between 1910 and 1911,79 corresponding closely to the capital cost of rebuilding Shed 17 — then the effect of the increased expenditure upon total costs must have been considerably cushioned by the spreading of the expense over the greater volume of revenue traff ic . (ii) The Incidence Of Disruption Available evidence permits a non-quantitative analysis of both the actual extent of traffic disruptions due to avalanches, and of the manner in which traffic disruptions were perceived by. 134 the C.P.R. and by the public. The importance of the manner in which the disruption was perceived should not be underestimated. When C.P.R. management undertook an investment solution to a problem, it was of course reacting to its perception of the problem. As the analysis in section (a) above revealed, the C.P.R. does not appear to have been dissatisfied with the perceived return which it obtained from its investment solutions to the avalanche problem. Neither does the press, insofar as it reflected, through its editorial and correspondence columns, the public's perception of traffic disruptions due to avalanches, appear to have been dissatisfied with the effectiveness of the C.P.R.'s avalanche-defence investments in reducing the experience of disruption in the Selkirks. In the wake of the 1910 disaster, the Vancouver Province reported that, It had almost become a byword that although occasional slides occurred the existence of snowsheds and a perfect system of patrolling the tracks near unprotected spots had hitherto, with rare exceptions, prevented any serious accident. No passenger or freight trains were ever swept away and no passenger ever lost his life.8 0 Although the 1910 disaster was followed by a succession of slides which interrupted traffic for almost two weeks,81 not all of these slides were in the Selkirks.82 Moreover, the C.P.R.'s handling of the disruption drew favourable press comment, for example from the Calgary Daily Herald: Now that the heaviest engagements of the trouble have been passed, the mountain staff are able to find in their achievement nothing but that which reflects creditably upon themselves...83 With passenger traffic forming a high proportion of total train movements through Rogers Pass,84 the C.P.R. must have been 135 acutely sensitive to the manner in which the public perceived the extent of traffic disruption due to avalanches. Yet it does not appear that the public was sufficiently alarmed for the C.P.R. to have been pressured by public opinion. The actual extent of traffic disruption may be established slightly more concretely than the perceived extent of disruption. The following analysis will first assess the extent of direct disruption to both passenger and freight traffic which was consequent upon avalanches, as reflected in data concerning train movements through the mountains. The analysis will then consider the efficacy of diversionary arrangements which were intended to palliate the disruption caused by snowslides. Finally, the analysis will attempt to determine whether the avalanche problem was increasing in severity in the years prior to the decision to abandon the surface alignment. In assessing the actual extent of disruption to passenger services, it must be conceded that there are at least four recorded incidents of passenger trains having been struck by avalanches in Rogers Pass prior to construction of the Connaught Tunnel. Two of these incidents occurred in the mid-1890's,85 one in January 1912,86 and one in April 1913,87 after the decision to abandon the surface alignment had been taken. No casualties were reported in any of the incidents. There were tales of miraculous escapes,88 and as soon as the 1911 slide season began, with the memory of the previous year's disaster presumably still fresh in the public's mind, the C.P.R. had to move rapidly to quash rumour of "a heavy snowslide at Rogers Pass."8' 136 Nevertheless, disruption to passenger services was generally in the form of delay rather than of physical damage or diversion: the standard operating procedure was to "hold" trains until it could be ascertained that the line was clear. Incidence of delays may have been quite frequent, but it cannot be proven that those delays were any more serious than delays caused by other operating problems encountered in providing the transcontinental service. The only complete record of passenger train performance through the Selkirks which is extant is that for 1908, and this is presented in table 2. The table records on a monthly basis the aggregate time gained and lost upon schedule of the daily transcontinental service whilst crossing the Mountain Subdivision of the C.P.R. main line. Train No. 96 was the eastbound service, or "Atlantic Express," which had a morning path across the Selkirks, departing from Revelstoke at 0830 and arriving in Donald at 1433. Train No. 97 was the westbound service, or "Pacific Express," which had an afternoon path, departing from Donald at 1405 and arriving in Revelstoke at 1925.'0 TABLE 2 PASSENGER TRAIN RECORD,' MOUNTAIN SUBDIVISION, 1908. Time Gained Time Lost Per Train Per Month Per Train Per Month (Hrs.-MinsTl (Hrs.-MinsTl Month Train No. 96 97 96 97 Jan. 1 -39 27-16 -40 -55 Feb. 7 -36 31-30 4 -10 -45 March 7 -32 30-28 5 -05 6 -04 April 7 -07 17-13 59 -07 54 -15 May 7 -25 11-09 n il 3 -13 June 7 -29 28-32 4 -15 -38 July 7 -45 26-26 4 -55 2 -32 Aug. 4 -37 33-29 5 -48 4 -58 Sept. 7 -35 37-46 1 -50 4 -58 Oct. 8 -35 11-40 11 -40 4 -20 Nov. 12 -51 1-20 13 -55 2 -30 Dec. 4 -49 17-12 2 -54 8 -51 Total 85 -00 280-01 124 -19 93 -54 Per Train Total Time Gained Per Year, All Trains: 365 hrs. 01 Total Time Lost Per Year, All Trains: 217 hrs. 13 mins. Source:- "Notebook," Kilpatrick MSS, Vancouver, n.p., n.d. 138 The table demonstrates that, in 1908 at least, there was as much potential for the service to recover time while crossing the Selkirks as there was for it to lose time. Losses incurred on the Mountain Subdivision averaged much less than an hour per day throughout the year, except in April, which appears to have been the peak month for slide activity in 1908. Even in April, however, losses still averaged less than two hours per day throughout the month. Except in April, the distribution of losses was not markedly skewed towards the winter months, but was generally uniform throughout the year. The fact that net gains outweighed net losses by some forty per cent, suggests that passenger trains were often already late when received onto the division. Performance data are available for all traffic on a monthly basis during the years 1906-08. The data for the average number of trains per day on both the Mountain and Shuswap sections, eastbound and westbound, are reproduced as table 3. From this table, it can be seen that, except in 1907, more trains per day were put through the Mountain Section than through the Shuswap Section during each period of the year. The number of trains per day shows no sharp change from month to month, although if snowslides had indeed imposed a constraint upon traffic movements through Rogers Pass during certain months of the year, then some discontinuity between the monthly totals might be expected. It may be inferred that the number of trains operated over the Mountain Section was determined by the availability of traffic rather than by the availability of paths between avalanches. When the demand for train movements was comparable 139 between the Mountain and Shuswap Sections, the Mountain Section could meet the demand at least as adequately as the Shuswap Section. The stochastic probability of any particular train being delayed on the Mountain Section was not sufficient to induce the C.P.R.. to run trains with less frequency over the Mountain Section than over the Shuswap Section. 140 TABLE 3 AVERAGE NUMBER OF TRAINS PER DAY, MOUNTAIN AND SHUSWAP SECTIONS, 1906-1908. Mountain Section Westbound Eastbound Total 1906 1907 1908 1906 1907 1908 1906 1907 1908 J 2.66 1.93 1.8 2.78 2.07 1.6 5.44 4 3.5 F 3.21 2.15 2.7 3.14 2.18 2.2 6.35 4.33 4.9 M 4.03 3.77 2.29 4.29 3.77 2.29 8.32 7.54 4.58 A 4.63 3.27 2.26 4.73 3.5 2.1 9.36 6.77 4.36 M 4 4.5 2.4 4.5 4.6 2.6 8.5 9.1 5 J 5.6 3.8 2.3 5.6 4.4 2.4 11.2 8.2 4.7 J 5 3.9 2.8 5 4 3.5 10 7.9 6.3 A 5 3.9 3.3 5.2 4 3.5 10.2 7.9 6.8 S 4.1 3.4 3.15 4.7 3.79 3.36 8.8 7.19 6.51 0 4.3 3.2 3.16 4.4 3.5 3.3 8.7 6.7 6.46 N 3.8 3.3 2.8 4 3.1 2.7 7.8 6.4 5.5 D 2.89 2.58 2.8 3.06 2.6 2.7 5.95 5.18 5.5 Average number of trains daily, . all year: 4.1 3.3 2.7 4.3 3.5 2.7 8.4 6.8 5.4 % +/- in all-year average number of trains daily, over previous year: -19.5 -18.2 -19.2 -22.0 -19.3 -20.1 Average number of trains daily, January-April: 3.6 2.8 2.3 3.7 2.9 2.1 7.4 5.7 4.3 % +/- in slide-season average number of trains daily, over previous year's slide season: -22.2 -17.9 -21.6 -27.6 -23.0 -24.6 141 TABLE 3 (Cont.) Shuswap Section Westbound Eastbound Total 1906 1907 1908 1906 1907 1908 1906 1907 1908 J 2.43 2.19 1.5 2.43 2.13 1.5 4.86 4.32 3 F 3.03 2.71 1.8 3.03 2.03 1.9 6.06 4.74 3.7 M 3.7 4.03 1.87 4 4 2 7.7 8.03 3.87 A 3.9 3.8 1.9 4.3 4.16 2 8.2 7.96 3.9 M 3.4 4.8 1.9 3.6 4.5 2 7 9.3 3.9 J 4.7 4 1.96 4.7 4.4 2.2 9.4 8.4 4.16 J 4.2 3.6 2.2 4.2 4.4 2.5 8.4 8 4.7 A 4.8 3.9 2.5 4.8 3.6 2.4 9.6 7.5 4.9 S 3.79 3.79 2.3 4.2 • 3.6 2.6 7.99 7.39 4.9 0 4 3.48 2.5 4.5 3.7 2.7 8.5 7.18 5.2 N 3.8 3.2 2.6 4 3 2.4 7.8 6.2 5 D 3.06 2.5 2.2 3.2 2.6 2.2 6.26 5.1 4.4 Average number of trains daily, all year • • 3.7 3.5 2.1 3.9 3.5 2.2 7.6 7.0 4.3 % +/- ir i all-year average number of trains daily, over previous year: -5.4 -40.0 -10.2 -37.3 -7.9 -38.7 Average number of trains daily, January-Apr i1: 3.3 3.2 1.8 3.4 3.1 1.9 6.7 6.3 3.6 % +/- in slide :-season average number of trains daily, over previous year's slide season: -3.0 -•43.8 -8.8 -38.7 -6.0 -42.9 Source:- "Notebook," Kilpatrick MSS, Vancouver, n.p., n.d. 142 Table 4 contains data for the average equivalent gross tonnage handled per trip on both the Mountain and Shuswap Sections, eastbound and westbound. Two features of this table should be noted. First, there is no sharp change from month to month in average train weights on the Mountain Section, although discontinuities might again be expected if snowslides were a constraint. Second, train weights on the Mountain Section were consistently less than train weights on the Shuswap Section, at all times of each year, except during the winter months of 1907. This suggests that the seasonal avalanche hazard did not exercise influence over the average weight of trains: the consistent difference between train weights on the two sections may be explained by the fact that the gradient system on the Mountain Section was more severe than that on the Shuswap Section. 143 TABLE 4 AVERAGE WEIGHT OF TRAINS, MOUNTAIN AND SHUSWAP SECTIONS, 1906-1908. (Equivalent Gross Tons Per Trip) Mountain Section Westbound "Eastbound 1906 1907 %+/- 1908 %+/- 1906 1907 %+/- 1908 %+/-J 618 509 -18 802 + 58 571 658 +15 739 + 12 F 565 677 + 20 749 + 11 610 592 -3 723 + 22 M 539 635 + 18 785 + 24 658 727 + 10 787 +8 A 554 677 + 22 820 + 21 665 688 + 3 798 +16 M 500 657 + 31 684 + 4 658 720 +10 830 +15 J 481 657 + 37 777 + 18 671 732 + 10 837 +14 J 495 621 + 25 775 + 25 671 755 + 13 828 +10 A 497 677 + 36 607 -10 685 741 + 8 734 -1 S 538 683 + 27 667 -2 704 735 +4 797 + 8 0 557 668 + 20 699 + 5 712 748 + 5 806 + 8 N 582 790 + 36 809 + 2 710 734 + 3 802 + 9 D 530 765 + 44 746 -2 695 716 + 3 764 + 7 Average Train-•weight Per Trip: 539 668 + 24 743 + 11 668 712 + 7 787 +11 TABLE 4 Shuswap Westbound 1906 1907 %+/- 1908 %+/-J 631 498 -21 835 + 68 F 541 563 + 4 907 + 61 M 559 603 + 8 882 +46 A 626 638 + 2 911 +43 M 530 687 + 30 844 + 23 J 509 634 + 25 938 + 48 J 527 669 + 27 916 + 37 A 491 681 + 39 760 + 12 S 515 675 + 31 753 + 12 0 546 619 + 13 800 + 29 N 575 765 + 33 897 + 17 D 513 714 + 39 958 + 34 Average Train-weight Per Trip: 547 647 +18 867 +34 144 (Cont. ) Section Eastbound 1906 1907 %+/- 1908 %+/-659 630 -4 732 +16 700 633 -10 760 + 20 712 730 + 3 833 + 14 731 709 -3 918 + 29 721 741 + 3 1030 + 39 720 728 + 1 977 + 34 735 736 - 1005 + 37 718 757 + 5 966 + 28 719 768 + 7 984 + 28 726 751 + 3 979 + 30 721 750 + 4 982 + 31 688 773 + 12 957 + 24 712 725 + 2 927 + 28 Source:- "Notebook," Kilpatrick MSS, Vancouver, n.p., n.d. 145 Table 5 records the total equivalent gross tonnage per month which was transported in either- direction over the Mountain and Shuswap Sections during the years 1906-08. Again, there are no systematic discontinuities between the monthly totals on the Mountain Section, as there might have been if snowslides had been a constraint. It may be inferred that the monthly variation in tonnages on the Mountain Section, and the monthly differences between the tonnages on the Mountain Section and those on the Shuswap Section, were due to the varying availability of traffic, that is, to varying demand, rather than to the varying availability of train paths during the avalanche season. TABLE 5 TOTAL EQUIVALENT GROSS TONNAGE PER MONTH, MOUNTAIN AND SHUSWAP SECTIONS, 1906-1908. Mountain Sect ion Westbound 1906 1907 %+/- 1908 %+/-in in Total Total Jan. 50,960 30,453 44,752 Feb. 50,782 40,755 58,647 March 67,337 74,212 55,727 April 76,951 66,414 55,596 May 62,000 91,652 50,890 June 80,808 74,898 53,613 July 76,725 75,079 67,270 Aug. 77,035 81,849 62,096 Sept. 66,174 69,666 63,032 Oct. 74,248 66,266 68,474 Nov. 66,348 78,210 67,956 Dec. 47,483 61,185 64,753 Total 796,851 810,639 + 2 712,806 Eastbound Jan-. 49,209 42,224 36,654 Feb. 53,631 36,136 46,127 March 87,507 84,964 55,869 April 94,365 72,240 50,274 May 91,791 102,672 66,898 June 112,728 96,624 60,264 July 104,005 93,620 89,838 Aug. 110,422 91,884 79,639 Sept. 99,264 61,520 80,338 Oct. 97,117 81,158 82,454 Nov. 85,200 68,262 64,962 Dec . 65,928 57,710 63,947 Total 1,051,166 889,014 -15 777,264 Year Total Tonnage %+/- Over Previous Year 1906 1,848,017 1907 1,699,653 -8 1908 1,490,070 -12 147 TABLE 5 (Cont.) Shuswap Section Westbound 1906 1907 %+/- 1908 %+/-in in Total Total Jan. 47,533 33,809 38,828 Feb. 45,898 42,720 47,345 March 64,117 75,333 51,130 April 73,242 72,732 51,927 May 55,862 102,226 49,712 June 71,769 76,080 55,154 July 68,615 74,660 62,471 Aug. 73,061 82,333 58,900 Sept. 58,556 76,748 51,957 Oct. 67,704 66,778 62,000 Nov. 65,550 73,440 69,966 Dec. 48,663 55,335 65,336 Total 740,570 832,194 +12 664,726 Eastbound Jan. 49,642 41,599 34,038 Feb. 59,388 35,980 41,876 March 88,288 90,520 51,646 April 94,299 88,483 55,080 May 80,464 103,370 63,860 June 101,520 96,096 64,482 July 95,697 100,390 77,888 Aug. 106,838 84,481 71,870 Sept. 90,594 82,944 76,752 Oct. 101,277 86,140 81,942 Nov. 86,520 67,500 70,704 Dec. 68,250 62,304 65,267 Total 1,022,777 939,807 -8 755,405 Year Total Tonnage %+/- Over Previous Year 1906 1,763,347 1907 1,772,001 +0.5 1908 1,420,131 -20 Source:- Tables 3 and 4. Average no. of trains per day x days per month x equivalent gross tonnage per trip. (1908 leap year). 148 In table 6, tonnages moved during the slide seasons (January to April) are compared with tonnages moved on an annual basis over the Mountain and Shuswap Sections. From this table, it appears that in the first third of each of the years 1906-08, slightly less than one third of the annual total tonnage of traffic was conducted over the Mountain Section. A similar proportion of total annual business was conducted over the Shuswap Section during the first four months of the year. In absolute terms, greater tonnages were conveyed over the Mountain Section than over the Shuswap Section during the slide seasons of two of the three sample years. Moreover, although traffic volumes over both sections were generally declining throughout the three-year period, column (vi) of table 6 suggests that the rate of decline on the Shuswap Section was faster during the slide season than on an annual basis. The difference between slide-season and annual rates of decline was not as great on the Mountain Section, and in 1908, indeed, the rate of decline was less during the slide season than the average annual rate of decline. TABLE 6 COMPARISON OF TRAFFIC ON MOUNTAIN AND SHUSWAP SECTIONS, SLIDE SEASONS (JANUARY-APRIL), 1906-1908. Mountain Section (i) Year 1906 1907 1908 (ii) Tonnage Moved 530,741 447,398 403,646 (iii) Annual Jan.-April Tonnage 28.7 26.3 27.1 (iv) %+/- Over  Jan.-April of Previous Year -15.7 -9.8 (v) %+/- Over Annual  Tonnage Of  Previous Year -8 -12 (vi) (iv)-(v) -7.7 + 2.2 Shuswap Section 1906 522,407 29.6 1907 481,176 27.2 -7.9 +0.5 -8.4 1908 371,870 26.2 -22.7 -20 -2.7 % Mountain Tonnage Greater % Annual Total Mountain Than Shuswap Tonnage During Tonnage Greater Than Slide Seasons Annual Total Shuswap 1906 1.6 4.8 1907 -7.5 -4.3 1908 8.5 4.9 Source:- Tables 3, 4 and 5. 150 The data must be treated circumspectly, for several trend factors operated concurrently across the period. It is difficult to determine whether the C.P.R. countered the risk of traffic disruption due to avalanches between January and April by running fewer but heavier trains. Again, it is difficult to determine whether gross tonnages transported between January and April were low because traffic was disrupted by avalanches during these months, or because the traffic, for example lumber,91 was seasonal in character. Indeed, it may be possible to explain some of the seasonality of the traffic, not only in terms of its composition, but in terms of the expectation of avalanche disruption per se. Monthly data are available for the aggregate gross ton mileage of all train movements over the Mountain Subdivision for the years 1910 and 1911. These data are particularly valuable in analysing the extent of disruption due to avalanches, as they span the slide seasons which several authorities believe to have been crucial in motivating the decision to abandon the Rogers Pass route. The data embrace the period of the 1910 avalanche disaster, and the two weeks of disturbance which followed. They also extend to within months of the decision to survey an underground route through the Selkirks, and include the penultimate winter of surface operation, one of the winters which, Lamb argues, "showed that slide hazards continued to be high," and persuaded the C.P.R. "that drastic action was essential."92 The data, aggregating the gross ton mileage of all passenger and freight trains, are presented in table 7, and a 151 comparison of train performance between the years 1910 and 1911 is presented in table 8. The caveats expressed in the analysis of the 1906-08 data are equally applicable to analysis of the 1910-11 data. Again, however, certain inferences may be drawn. 152 TABLE 7 TOTAL EQUIVALENT GROSS TON MILEAGE PER MONTH, MOUNTAIN SUBDIVISION, 1910 AND 1911. 1910 1911 Jan. Feb. March April May June* July Aug. Sept. Oct. Nov. Dec. 53,348,399 50,058,868 60,162,544 72,008,888 79,289,709 86,061,150 85,431,095 83,407,130 78,025,943 80,259,054 69,494,910 64,499,788 48,946,968 60,621,556 81,423,638 89,572,739 98,883,576 92,546,832 99,277,235 102,947,877 95,277,584 95,710,487 77,880,170 77,035,828 Total 862,047,978 1,019,124,490 * No record of June ton mileages in RCM. Ton-mileage figures shown have been calculated from available figures for Total Train Miles in June, multiplied by estimate of average train weights in June. Average train weights for June are averages of train weights for May, July and August of respective years. Source:- C.P.R. Co., "Statement of Train Locomotive and Actual Gross Ton Mileage, First District, British Columbia Division," Form S.O. 46, RCM. TABLE 8 COMPARISON OF EQUIVALENT GROSS TON MILEAGES, MOUNTAIN SUBDIVISION, 1910 AND 1911. 1910 Jan-April Total Annual Gross Ton  Mi leage 235,578,699 862,047,978 % Of Annual Total  Conveyed Jan-April 27.3 1911 Jan-April Total Annual 280,494,901 1,019,124,490 27.5 %+/- Between Jan-April 1910 and Jan-April 1911: +19.1 %+/- In Total Annual Gross Ton Mileage, 1910-1911: +18.2 Source:- Table 7. 154 The proportions of total annual business which were conducted during the slide seasons of 1910 and 1911 were not markedly less than those of 1906-08. Indeed, the proportion of annual business carried during the slide season of 1911 may have been greater than the proportion for any year since 1906. (See table 6, column (iii) ). The absolute volume of business conducted during the slide season of 1911 increased by 19.1% over the corresponding period of 1910, slightly more than the increase in the absolute volume of business on an annual basis, which was 18.2%. During these two years of increasing traffic, the main line continued to conduct slightly less than one third of the total annual business during the first one third of the year. Nevertheless, the proportion of total annual business which was conducted during the slide season of 1911 did increase slightly over that conducted during the slide season of 1910, from 27.3% to 27.5%. This evidence suggests that, within one year of its decision to abandon the surface alignment, the C.P.R. could still meet additional demand, where required, by increasing the volume of traffic conveyed over the mountains between January and April. The evidence highlights the difficulty of convincingly apportioning influence upon the volume of business conducted in March 1910 between demand factors and avalanche factors. Despite the fortnight's interruption during March 1910, business increased by 20% over the previous February, and would increase by a further 20% in April. The volume of business conducted in March 1910 was greater than it had been for at least three months, greater than it would be in January 1911, and almost as 155 great as it would be in February 1911. However, it must be conceded that the volume of business which was conducted in March 1911 was over 35% greater than the volume of business conducted in the corresponding month of 1910. This increase between the March figures of 1910 and 1911 was almost double the annual increase over the respective years (35% against 18.2%). Moreover, in 1911, a less troubled season, traffic increased by 34% between February and March, and by only 10% between March and April. This suggests that the blockages of March 1910 may have retarded the rate of growth in traffic volume which could usually be expected to prevail between February and March. It suggests also that the relatively high rate of growth which prevailed between March and April 1910 may have been due to a "catching-up" process, in which a backlog of traffic was cleared as soon as mountain access was restored. Nevertheless, the very difficulty of convincingly allocating weightings between demand factors and avalanche factors in part refutes the thesis that the "disastrous" 1910 slide season was a decisive influence in motivating the abandonment of the surface alignment. This analysis of train movements, spanning years of both declining and increasing traffic volumes, does not establish conclusively that avalanche disruptions to traffic flows through the Selkirks seriously limited monthly traffic movements. The analysis does not reveal whether the relative increase in train movements which occurred during the slide seasons of certain years fully satisfied the demand for train movements in those seasons. Equally, however, the analysis does not prove that the C.P.R. would have been justified in abandoning its surface 156 alignment through Rogers Pass purely because of the extent of actual disruption to traffic from snowslides. Only if it could be proven, first that the demand for train movements through the Selkirks between January and April could not be met, and second that it could not be met because of interruptions from avalanches, could it be concluded that the extent of disruption to traffic which was caused by snowslides prompted the C.P.R. to undertake investment in an alternative route across the mountains. The evidence of this analysis offers no such proof. Instead, it proves that the demand could be met with no more difficulty in the Selkirks than in the Gold Range, and suggests that avalanches were not a decisive .factor in preventing the demand for traffic movement through the Selkirks from being satisfied. (iii) Diversionary Arrangements In assessing the efficacy of diversionary arrangements intended to mitigate the disruption due to avalanches, it must be remembered that until completion of the Canadian Northern main line through the Yellowhead Pass in 1915, there was no all-rail alternative to the C.P.R. main line through the Rockies and the Selkirks. Moreover, until 1914, the only alternative route which did exist entailed diversion through both the Rockies and the Selkirks, even when the C.P.R. main line was blocked through only one of these mountain ranges. For two weeks in March 1910, when the main line was blocked by slides in the Rockies and the Selkirks,'3 and again for at least three days in September 1913, when the main line was blocked by slides in the Rockies,'* all 157 traffic, "passenger, mail, baggage and express," had to be completely diverted between Calgary and Revelstoke. Westbound traffic was diverted via the Crow's Nest Pass railway to Kootenay Lake, conveyed by barge from Kootenay Landing to Nelson or Procter, thence by rail to Slocan, by barge across Lake Slocan, by rail to Nakusp, by barge again across the Upper Arrow Lake to Arrowhead, and thence by rail to Revelstoke. Eastbound traffic followed the route in reverse. Up to six transhipments were involved in either direction. It was not until 1914, with the opening of the Kootenay Central from Golden to Colvalli,95 linking the main line through the mountains with the Crow's Nest Pass line, that it became possible for the C.P.R. to divert traffic around the Selkirks without also having .to divert it around the Rockies. Transhipment at Kootenay Landing would continue to be necessary until 1930. 5 4 Diversionary arrangements were therefore cumbersome and expensive. In 1910, at least, they were also inadequate. The Revelstoke Mai1-Herald reported that, Freight sent round by the Crow's Nest has been much delayed owing to the barge equipment on Slocan Lake not being capable of handling the immense quantity of congested freight which had accumulated during the past two weeks. 9 7 The inadequacy of the diversionary arrangements, together with the "awful disaster at Rogers Pass" and "the danger to passengers," prompted the Mail-Herald to urge completion of the Arrowhead and Kootenay Railway, "so as to give the C.P.R. an alternative route over the mountains under any conditions."58 The question of the desirability of alternative routes to the C.P.R. main line through the Selkirks will be discussed in 158 chapter 7. It has already been noted that the Revelstoke Mail- herald was interested in the Arrowhead and Kootenay line chiefly for developmental reasons, and not because of its value as a safe alternative to the C.P.R. main line. The editor of the Mail-Herald claimed that, The immense sum which the present disaster, and the extensive slides which accompanied it, has cost the company, would complete the forty miles of the Arrowhead and Kootenay railway remaining to connect the main line with the Crows Nest road, and this would give the C.P.R. practically as short a route across the mountains as the main line, while on the Arrowhead and Kootenay railway there is not a snowslide...5' The C.P.R. clearly did not share his conviction: the Arrowhead and Kootenay was not completed until 1930. Diversionary arrangements were only required on those occasions when the avalanche defence system on the main line failed. The diversions of March 1910 and September 1913 are the only recorded occasions of such failure. Even if diversion was cumbersome, expensive and inadequate when required, evidence suggests that it was not required sufficiently frequently to justify the channelling of investment from the main-line avalanche defence system into the construction of an alternative route through the Selkirks. (iv) Was Disruption Increasing? In analysing the role of traffic disruption in motivating the decision to construct the Connaught Tunnel, it is important to determine whether the extent of disruption due to avalanches was increasing in the years immediately prior to the abandonment of Rogers Pass. Certain authorities maintain that it was avalanche experiences in these years which induced the C.P.R. to 159 discontinue surface operations.100 Moreover, contemporaries may have perceived that the threat from avalanches was increasing in intensity. In the wake of. the 1910 disaster, the editor of the Revelstoke Mai1-Herald wrote thus: ...we understand from railway men, who know what they are talking about, that these slides are getting worse every year, and will continue to do so, as the-remains of the forests which formerly protected a great part of the road, but were destroyed by fire during and since construction, rot away, thus giving clear sweep to slides from the deep snow banks which cover the mountains in winter.101 The previous analysis of train movements suggests that disruption did not increase throughout the period 1906-11. The proportion of total annual traffic which was conveyed through the Selkirks during the slide season remained almost constant at around 28%. This proportion actually increased when the total volume of traffic increased. Whilst this evidence cannot be considered conclusive, since it is impossible to discriminate the influence of seasonal demand factors from avalanche factors upon the volumes of traffic conveyed in various months of the year, further evidence is° consistent with the view that disruption did not increase, either with the passage of time, or with increasing traffic volumes. By February 1911, the C.P.R. had experienced its largest snowfall for sixteen years in the Selkirks.102 Nevertheless, in the Engineering Department's annual report for the year ending June 30, 1911, the Chief Engineer, Sullivan, reported that, "During the winter the only slide of importance was one at M.B. 93, Mountain Subdivision, which covered one end of the shed delaying traffic six hours."103 The corresponding report for 1912 noted: 160 Small snowslide on January 14th at Mileage 113, Albert Canyon, carried out three cars from wrecking train...killing one man. The same morning at Shed No. 11, Colonist Car No. 14 was cut out from the train, but no one was hurt.104 By this time, the decision had already been taken to double track through the Selkirks, and the way was paved for the abandonment of Rogers Pass. A list of avalanche occurrences in the Selkirks, recorded by the C.P.R. between 1909 and 1918 confirms that the years 1911 and 1912 were unremarkable for their slide activity, and that the extent of such activity would not be sufficient to cause disruption warranting investment in an alternative route. For 1911, the list records simply, "Nothing." Throughout the avalanche season of 1912, the list records two slides on January 14, and then only one more slide, over two months later, which apparently required only "1 hr. 30 min. to clear."105 Certainly, there was disruption to traffic in the years immediately before construction of the Connaught Tunnel, and this disruption was caused by slide activity. However, much of the slide activity did not occur in Rogers Pass at all. In November 1909, eastbound passenger train No. 2 was cut off by slides at front arid rear in the Fraser Canyon,106 and the main line was blocked for two days between Barnet and Lytton, whilst in the Selkirks and the Rockies, "the C.P.R. sustained no damage."107 In January 1911, the main line was closed for a week by slides east of Field.108 In April 1912, a passenger train was struck by a slide at Savona's Ferry, causing the only recorded avalanche-induced casualties in passenger service over the entire C.P.R. main line through the mountains prior to 161 construction of the Connaught Tunnel. Even then, the fatalities were traincrew, not passengers.109 In August 1913, the C.P.R. experienced a "series of blockades... in the mountain sections" which was "unparalleled in the history of the company. Never before have such a number of slides come down so many different points."110 Again, however, most of the disruption was in the Kicking Horse Canyon, which would be rendered no less vulnerable to disruption from slides by the construction of a tunnel beneath the Selkirk Mountains. The results of this analysis suggest that the extent of disruption to traffic due to avalanches may not actually have been very great, and was unlikely to have been sufficient to justify abandonment of the surface operation over the Selkirks. The extent of disruption does not appear to have warranted investment in diversionary facilities as an alternative to the route through Rogers Pass. Neither does the extent of disruption appear to have increased, either with the passage of time, or with increasing traffic volumes. (v) Disruption Costs And The Abandonment Decision Since this analysis makes no attempt to quantify the cost of disruptions to traffic due to avalanches, it is not possible to assess the relative importance of disruption costs in motivating the decision to construct the Connaught Tunnel. It is possible, however, to estimate how high those disruption costs would have had to have been in order to justify investment in a tunnel. The analysis of the direct cost of maintaining the 162 avalanche defence system revealed that the potential benefit of discontinuing the system, $3,600,200, was outweighed by the potential cost of tunnelling, at least $5,495,000. Cost outweighed benefit by some $1,894,800. If the benefits of tunnelling were to have outweighed the costs, the potential benefit of avoiding disruption to traffic from avalanches, combined with the non-quantified direct cost of maintaining the avalanche defence system, would have had to have been at least $1,894,800. Discounted at four per cent., these non-quantified costs would have had to have exceeded $75,000 per year,111 in addition to the $125,000 per year expended in maintaining the avalanche defence system. If the non-quantified benefits were expected to be so high — sixty per cent, of the quantified benefits of discontinuing the avalanche defence system — it is reasonable to suppose that the C.P.R. would have included them in an evaluation of the forecasted cost savings of constructing a double-track tunnel which would obviate disruption from avalanches. The Company published an evaluation of the project in 1914,112 which sought to justify construction of a double-track tunnel entirely by reference to the magnitude of the cost savings of the project, without any explanation of the costs which would have to be incurred in order to obtain those savings. Yet even with this bias, the evaluation makes no allusion to anticipated benefits from reduced traffic disruption. One interpretation of this omission might be that the magnitude of savings which could be derived from a reduction in traffic disruption was not sufficiently great to influence the decision of whether or not to construct a tunnel beneath 163 Rogers Pass. This analysis of snow problems in the Selkirks raises serious doubts about the adequacy of the explanation that the surface alignment through Rogers Pass was abandoned because of the severity and the intractability of the avalanche hazard. Neither the 1910 disaster in particular, nor the snowslide problem in general, with its concomitant costs in defence and disruption, appears to have motivated the decision to construct the Connaught Tunnel. Nevertheless, the view that the tunnel was constructed in response to avalanche problems remains thoroughly entrenched. Any of three reasons may explain its endurance. The first may be that the cost-savings approach by which the C.P.R. sought to justify the tunnel investment in retrospect accorded paramount importance in the outcome of the evaluation to the role of savings in the avalanche defence system. Thus: The figures would not have been very decisive one way or the other were it not for the fact that there is now 4 1/2 miles of wooden snow sheds on the present location which will all be done away with on the new location. The maintenance and cost of renewals of these sheds cost between $85,000 and $100,000 per year. To maintain and renew a double track wooden shed would probably cost at least 50% more than the above, so that with a saving of about $125,000 per year in maintenance and renewals of snow sheds and a calculated saving in operation and maintenance of $171,271.22 on a traffic that surely will be reached in the near future, there was no doubt as to the proper course to pursue.113 If the anticipated savings in snowshed costs had indeed constituted such a high proportion of the benefits of the project — almost three-quarters of the benefit which was expected to accrue from operating savings — then it might reasonably be asserted that the savings in snowshed maintenance 164 did tip the balance in favour of the abandonment of Rogers Pass. However, as will be explained more fully in chapter 8, the evaluation which the C.P.R. published omitted all consideration of congestion costs and of the opportunity costs of traffic which would have to be foregone if the single line over Rogers Pass was not doubled. These omissions lend a downward bias to the estimate of operating savings accruing from the project, and this in turn lends an upward bias to the contribution of savings in snowshed maintenance towards the total benefits of the tunnel. The second reason may be that Sir George Bury, who was instrumental in the decision to tunnel beneath Rogers Pass, himself emphasised in retrospective correspondence the role of avalanches in motivating the decision. Thus, two years after construction of the tunnel was commenced, Bury wrote, The big problem we had, of course, was to get the tunnel built as fast as possible, not only to avoid the continual expense of renewing the present snow sheds, and pay interest and overhead charges on a long delayed job, but to get away from the dangers of snow which is [sic] greater or less, depending on the season.114 As will be explained more fully in chapter 7, however, Bury himself offered a different rationale at the time when the decision was actually made. In June 1913, he wrote quite simply that, Regarding the necessity of this tunnel, I may say that the traffic over Rogers Pass has become so great that it is necessary to double track the line.115 He then proceeded to calculate an annual saving in traffic operating costs from the tunnel of $150,000 within four years. The final reason is the obvious one, that savings in both snowshed maintenance and disruption to traffic from snowslides 165 did accompany construction of the tunnel, since rail operations ceased to be conducted on the exposed surface of Rogers Pass. The analysis in this chapter indicates that snow problems in the Selkirks were a matter of ongoing concern to the C.P.R., but that they were not a crucial factor in motivating the abandonment decision. As long as the service was operated over Rogers Pass, the potential for disaster was always present. Both the C.P.R. and the general public were well-informed about the nature of the avalanche hazard, and were prepared to accept the risks in order to maintain the service. Routinely, the service was maintained without incident. As the Calgary Daily Herald obs'erved in the winter of 1910, Snowslides are looked for in season, as a matter of course, and under ordinary conditions they are attended with no special danger or delay to traffic. In these regions the battle with the snow is a commonplace feature of the winter's work. The organization and equipment are prepared for it.116 Occasionally, the potential for disaster was realised. However, the mere incidence of disaster would not necessarily precipitate a major change in policy: it would not necessarily dictate the cessation of rail operations through Rogers Pass. Whether or not the incidence of disaster would precipitate a major policy change would depend not upon the magnitude of the disaster itself, but upon the crucial question of how much the public or the railway company would be prepared to spend in order to avoid a repetition of the disaster. This was clearly perceived and understood by contemporaries: at the second inquest into the 1910 disaster, the C.P.R.'s Resident Engineer in Revelstoke asserted that, Tc make the road absolutely safe from slides, we 166 would have to have a continuous shed through the mountains — which is impracticable.117 The evidence of this analysis suggests that the C.P.R. was not prepared to spend the amount which would be required for a tunnel beneath the Selkirks, merely in order to avoid either the actual cost of operating services through hazardous avalanche paths, or the potential for disaster. Indeed, the analysis indicates that the C.P.R. should not have been prepared to spend that amount, for the benefits which it could expect to obtain from avoiding actual costs and potential disasters would not have offered the Company a return on its investment. Yet the C.P.R. did invest in a tunnel beneath the Selkirks, and when that tunnel opened, the Company did cease operating services through Rogers Pass. Clearly, if the benefits to be derived from avoiding avalanche problems did not alone justify investment in a tunnel, then the project must have been undertaken in the expectation of benefits from another source. Moreover, these latter benefits, aggregated with the former, must have been expected to justify the cessation of surface operations through Rogers Pass, operations which had been conducted for almost thirty years. The remaining chapters of this thesis will examine the nature and extent of these other benefits, and their influence upon the realignment schemes which were proposed. 167 FOOTNOTES 1 See, for example, P. Berton, op. cit., p. 335; P. Mason, "89 Over The Top," Canadian Rail, No. 257, June, 1973, p. 175; Snow  War, op.cit.,p.10. 2 There is some controversy over the actual number of victims. It was initially feared that over one hundred railwaymen had been buried. T. Kilpatrick, tel. to G. T. Bury, March 5, 1910. Revelstoke City Museum, (henceforth "RCM,") File No. 76.15.171-9223.2. However, the Vancouver Province of March 5, 1910 reported 61 dead, and the next day, the Calgary Daily Herald reported that, "All the men in the section gangs in that vicinity have been checked up and the total number missing is 62. There is no question that this is the total death list." The toll reported to the Board of Railway Commissioners was 58. Board of Railway Commissioners, Annual Report, 1911. DSP Vol. XLVI , 1912, 20c, p. 45. The last body was not recovered until over a month after the disaster. T. Kilpatrick, tel. to G. T. Bury, April 18, 1910. RCM, ibid. 3 Toronto Globe, March 7, 1910, p. 1. 4 Province, June 22, 1907, p. 1. 5 Dept. of the Attorney General, "Inquisitions, 1862-1918," No. 72, "James Moffat and Others, March 14, 1910," p. 10. PABC. 6 Kilpatrick testified that, "The diversion east of Rogers Pass has been very expensive, as extra precaution had been taken against slides." Revelstoke Mail-Herald, March 12, 1910, p. 1. 7 Province, March 8, 1910, p. 1. 8 Revelstoke Mail-Herald, March 16, 1910, p. 3. ' Province, March 8, 1910, p. 1. 10 Toronto Globe, March 2, 3, and 4, 1910. 11 "Inquisitions," op. cit., p. 19. 12 Ibid., p. 20. 13 Ibid., p. 8. 14 Ibid., p. 15. 15 Calgary Daily Herald, March 14, 1910, p. 1. 16 "Inquisitions," op. cit., p. 5. Ibid., p. 24. 168 16 Ibid., p. 10. 19 Ibid., p. 1. 20 Abbott to Van Home, January 4, 1888. PIC, CPCA. Either the instructions were ignored, or they lapsed with time, for a bridge carpenter at the second inquest into the 1910 disaster testified that, "We were doing just the same as we had always done, in working on slides at night." "Inquisitions," op. cit., p. 25. Only one employee was killed in the 1888 incident, although Robert Marpole, later to become General Superintendent himself, had a narrow escape. 21 On March 7, 1910, the B.C. Legislative Assembly carried a motion of sympathy for the victims of the "deplorable accident." Journals Of The Legislative Assembly Of The Province Of British  Columbia, Victoria, Vol. XXXIX, 1910, p. 91. Clearly, they could not be expected at this time to prejudice the outcome of the forthcoming inquest. 22 DSP, Vol. XLV, 1912, 20c, p. 45. 23 Labour Gazette, Journal of the Department of Labour, Ottawa, Vol. 10, April, 1910, p. 1177. 24 HoC Debates, Vol. XCIX, "Accidents on Canadian Railroads," February 20, 1911, pp. 3922-47. 25 Revelstoke Mail-Herald, March 16, 1910, p. 3. 26 Calgary Daily Herald, March 7, 1910, p. 1. 27 Revelstoke Mail-Herald, March 12, 1910, p. 4. 28 "Inquisitions," op. cit., pp. 19-20. 29 For example, on June 17, 1911, the C.P.R. paid $500 to Fredrik Vilhelm Carlsson as compensation for the death of his brother, Vic. "It is distinctly understood, however, that the Company does not admit but. disputes legal liability for the accident, the amount being given as a gratuity and for the sake of peace." RCM 76.15.165-91. 30 Shaughnessy to Bury, March 15, 1910. Letterbooks, op. cit. 31 Revelstoke Mail-Herald, March 12, 1910, op. cit. The Resident Engineer, J. P. Ford, stated that the flat land was "on south side." At this point, however, the main line is running in a north-south direction, and the Engineer means "east." 32 "Inquisitions," op. cit., p. 17. 33 Reported in, Bury to Shaughnessy, March 18, 1910. PIC CPCA. 34 RCM 76.15.165-57. 169 35 $667.02 was paid to the Manitoba Bridge and Iron Works Company in April, 1911. Ibid., 76.15.139-64. 36 Bury to Shaughnessy, March 15, 1910, PIC CPCA. 37 Revelstoke Mail-Herald, April 30, 1910, p. 9. 38 Province, March 8, 1910, p. 1. 35 Whyte to Shaughnessy, April 13, 1911, PIC CPCA. 40 "Inquisitions," op. cit., p. 15. 41 Revelstoke Mail-Herald, April 13, 1910, p. 9. 42 For further discussion, see below, Chapters 5.2 and 8. 43 "Draft of letter from Walter Moberly re. double-tracking of the C.P.R. west of Winnipeg, and the danger of Rogers Pass." n.d., Moberly MSS, p. D946. 44 Victoria Daily Times, July 6, 1910, p. 14. Earlier in the year, an editorial in the Revelstoke Mai1-Herald commented that, "While everybody is glad to see the initial steps taken to open up the Big Bend by a railway, the policy of tieing [sic] up the whole of the Canoe and Columbia valleys between Revelstoke and the Yellow Head Pass is all nonsense, and it is to be hoped these reservations will not be kept long in force." Revelstoke  Mail-Herald, April 30, 1910, p. 4. 45 Revelstoke Mail-Herald, June 11, 1910, p. 1. 46 Revelstoke Mail-Herald, April 13, 1910, p. 4. 47 See, for example, Lamb, op. cit., p. 265; O. S. A. Lavallee, "Rogers Pass: Railway to Roadway," Canadian Rail, op. cit., p. 155; J. A. Beatty, "CP Rail's Connaught Tunnel," Canadian Rail, No. 271, August 1974, p. 227; Pugsley, op. cit., p. 47. 48 Lavallee, ibid. 45 By June, 1888. See Chapter 4. 50 Statement of Snowsheds, "August, 1898," 6 pages, Kilpatrick Add Mss 323, PABC. 51 Ibid. 52 "C.P.R. Co. District No. 1, Pacific Division, Length of Snowsheds, October 8th 1904," ibid. 53 Assuming that the maintenance cost of 23,760 feet of snowshedding was $85,000-$100,000 per annum. See Chapter 7. 54 The C.P.R.'s plan of the relocation, dated October 19, 1908, shows that Sheds 15, 15A and 16, 426 feet, 89 feet and 353 feet 170 long respectively, were rebuilt; a new shed, 750 feet long, was required between Sheds 15A and 16, and the old Shed 17, 3,101 feet long, was replaced by a shed only 127 feet long. PABC, Maps Division, R-R 1, #9054. 5 5 See note (53). 56 C.P.R. Co., "Comparative Statement of Expenses, 1911 and 1910, First District, B.C. Division," RCM 76.15.140. 57 Ibid. 58 For the months of March, and May-December, 1912. C.P.R. Co., "B.C. Bridge and Building Dept., Distribution of Labour and Material," RCM 76.15.135-23; 76.15.128-34; 76.15.158-25; 76.15.130-25; 76.15.138-84; 76.15.131-30; 76.15.125-26; 76.15.167-78; 76.15.123-29. 5 9 See Chapter 7. 60 Sullivan to Bury, op. cit. 61 Ibid. 62 Ibid. 63 Ibid. 64 Ibid. 65 ($125,000/4%) + $475,200. 64 Sullivan to Bury, op. cit. '7 See, for example, C.P.R. Co., "B.C. Bridge and Building Dept.," op. cit., RCM 76.15.135-23. 68 See Chapter 4, note (126). '9 "B.C. Bridge and Building Dept.," op. cit. 70 Ibid., June, 1912. RCM 76.15.158-25. The expenditure upon "Watching" was $619.03. 71 "Appropriations 1902-3 (Field Book 360)," Kilpatrick MSS, Vancouver, p. 42. 72 Canadian Railway And Marine World, Toronto, April, 1911, No. 158, p. 323. 73 Bury to Shaughnessy, March 15, 1910, op. cit., PIC CPCA. 74 See above, pp. 92-93. 75 $5,495,000 - $3,600,200. See note (65). 171 76 In addition to the daily transcontinental passenger trains, two eastbound freight trains were scheduled, at least until November. Donald Truth, November 3, 1888, p. 3. 77 In 1908, 7% of the total annual number of trains crossed the Mountain Section during March. See table 3. In 1912-13, the total annual number of trains through Rogers Pass was 6,162. Cornell Civil Engineer, op. cit., p. 81. 7% of this total would have represented 14.42 trains per day during a 31-day month. See table 11. 79 See above, note (56). 80 Province, March 5, 1910, p. 1. See also, Calgary Dai1y  Herald, March 7, 1910, p. 1. 81 Disruption continued until at least March 15. Calgary Daily  Herald, March 15, 1910, p. 1. 82 Slides occurred at Field, Palliser and Glenogle in the Rockies, and at Three Valley in Eagle Pass. Revelstoke Mail- Herald, March 9, 1910, p. 8. One of the slides was 100 yards long and 50 or 60 feet deep. Calgary Daily Herald, March 7, 1910, p. 1. 83 Calgary Daily Herald, March 12, 1910, p. 9. 84 See below, pp. 191-195. 85 Tait, Memoranda to Shaughnessy, February 20, 1895 and February 29, 1896. PIC CPCA. 86 Bury to Shaughnessy, "Engineering Department, Annual Report for year ending June 30, 1912," September 7, 1912, PIC CPCA. 87 J. M. McKay, Superintendent, Revelstoke, to W. H. D'Arcy, General Claims Agent, Winnipeg, February 7, 1914, RCM 76.15.100-014872. 88 See, for example, Revelstoke Mail-Herald, March 9, 1910, p. 8. 89 Province, January 12, 1911, p. 1. 90 C.P.R. Co., "Annotated Timetable, corrected to January 10, 1908," Montreal. 91 See Chapter 6. 92 Lamb, op. cit., p. 265. 93 Province, March 9, 1910, p. 1. 94 Province, September 8, 1913, p. 15. 172 95 Innis, op. cit., pp. 154-155. 94 Lamb, op. cit., p. 212. 97 Revelstoke Mail-Herald, March 1.6, 1910, p. 3. 98 Revelstoke Mail-Herald, March 12, 1910, p. 4. 99 Ibid. 100 Lamb, op. cit., p. 265; Frontier Guide to the Incredible  Rogers Pass, Frontier Book No. 8, Calgary, 1963, p. 42. 101 Revelstoke Mail-Herald, March 12, 1910, p. 4. 102 Whyte to Shaughnessy, February 8, 1911, PIC CPCA. 103 Whyte to Shaughnessy, "Engineering Department Annual Report for year ending June 30, 1911," September 7, 1911, PIC CPCA. 104 Bury to Shaughnessy, "Engineering Department Annual Report for year ending June 30, 1912," op. cit. 105 "Avalanche Occurrences 1909-1918, as recorded by the C.P.R.," Mount Revelstoke and Glacier National Parks, Revelstoke, B.C., File No. 1779. xo£ Victoria Daily Colonist, November 30, 1909, p. 1. 107 Province, November 29, 1909, p. 1. 108 Province, January 11, 1911, p. 2; January 12, 1911, p. 1. 109 Victoria Daily Times, April 11, 1912, p. 1. 110 Province, September 8, 1913, p. 15. 111 $1,894,800 x 0.04 = $75,792 per year. 112 Cornell Civil Engineer, December, 1914, op. cit. 113 Ibid., p. 84. 114 Bury, Memorandum to Shaughnessy, July 23, 1915, PIC CPCA. 115 Bury to Shaughnessy, June 16, 1913, PIC CPCA. 116 Calgary Daily Herald, March 12, 1910, p. 1. 117 "Inquisitions," op. cit., p. 16. 173 CHAPTER 6 CAPACITY PROBLEMS After the turn of the century, there was a considerable increase in the volume of traffic requiring haulage over the Selkirks. For railways traversing mountainous regions, substantial increases in traffic volumes may entail significant changes in operating procedures, and major investments intended to obtain increments in line capacity. At issue in Rogers Pass, then, is the question of whether main-line capacity was adequate to meet the requirements of burgeoning traffic in the early years of the 20th Century. The analysis begins with a consideration of the capacity of the main line through Rogers Pass. Then it considers changes in the traffic demands which were imposed upon the line, and changes in the competitive pressures which the C.P.R. confronted in meeting these demands through the mountains of B.C. Next, improvements which the C.P.R. undertook on its system in response to these changes are analysed, and the financial resources of the C.P.R. to undertake further system improvements are appraised. Finally, forecasts of traffic flows through the mountains are examined, and the implications of these forecasts for main-line operations through the mountains are assessed. 6.1 The Capacity Of The Main Line The capacity of a railway line is the weight of traffic which can be transported over the line per unit of time. There are two crucial determinants of line capacity. The first is 174 train weight; that is, the maximum weight of traffic which can be conveyed over the line by a single train. The second is the number of train paths which the line can supply; that is, the maximum number of trains which can be passed over the line in a given unit of time. a) Train Weight As was explained in chapter 2, the maximum weight of trains which can be conveyed over a line segment or a division is determined by the ruling gradient of that line segment or division. Over the line segment between Beavermouth and Rogers Pass on the east slope of the Selkirks, the ruling gradient against westbound traffic was 2.2% compensated for 20.8 miles. Over the line segment between Albert Canyon and Rogers Pass on the west slope, the ruling gradient against eastbound traffic was 2.2% compensated for 25.3 miles.1 Table 9, containing tonnage ratings for a single locomotive through the Selkirk Mountains from Revelstoke to Field, demonstrates the influence of these gradients upon the weight of traffic which could be hauled in a single train through Rogers Pass. With the motive power which was available in 1913, any train travelling in either direction with an equivalent gross weight exceeding 508 tons required the attachment of a pusher locomotive for the entire distance of the ascent.2 TABLE 9 TONNAGE RATINGS FOR SINGLE 210% LOCOMOTIVE BETWEEN STATIONS ON MOUNTAIN SUBDIVISION, PRIOR TO JUNE 1913. Westbound Station Miles Eastbound Rating (tons) Rating (tons) Field Downgrade Ottertail 8. 2 1,224 Downgrade Wapta 4. 4 2,100 1,398 Leanchoil 4. 4 2,100 Downgrade Golden 18. 2 554 2,100 Donald 16. 3 2,100 2,100 Beavermouth 11. 7 1,283 508 Rogers Pass 20. 8 Downgrade Downgrade Albert Canyon 25. 3 508 Downgrade Revelstoke 20. 9 1,108 Source:- F. J. Fisher to J. M. McKay, Superintendent, Revelstoke, May 25, 1913. RCM 76.15.188-45693.6. 176 Even with a pusher, however, the maximum weight of a freight train on either side of the summit was restricted to 1,016 tons, still significantly less than could be hauled on adjoining sections through the mountains. More than one pusher was rarely attached to a train. It is not known whether this was because the C.P.R. was chronically short of locomotives with which to supplement its pusher fleet;3 whether it was because the incremental drawbar stress imposed by the additional locomotives would have fractured the car-drawbars;4 or whether it was because of the difficulty of co-ordinating more than two steam locomotives. Whatever the reason, the fact that only two .locomotives were generally allocated to any single train meant that trains exceeding 1,016 tons in equivalent gross weight had to be "cut" to this weight and forwarded in sections. In turn, this decreased the payload per locomotive-movement, and increased the number of train paths required for the movement of the traffic. Moreover, it set a ceiling on the extent to which an increase in the total volume of traffic requiring transit through the mountains could be accommodated simply by increasing the weight of trains. An increase in total volume above this ceiling would require more trains, and proportionately more locomotives, simply in order to match the capacity of adjoining sections. More trains, on a single line, and with the need to return pushers light through the increased opposing traffic, would, beyond a certain point, impose congestion costs additional to the direct operating costs of the train movement. The addition of a pusher locomotive, and the addition of an entire train, both represent "lumpy" investments. That is, the 177 investment produces the same fixed increment in capacity regardless of whether there are five tons or five hundred tons of additional traffic to be moved. Table 10 presents average train weights through the Selkirks for the period 1906-13. In the years 1906-08, "Consolidation" locomotives were deployed as pushers in the Selkirks. Since these locomotives had a rating of 490 tons,5 the fixed increment in capacity which was obtained from the addition of one pusher was 490 tons. The maximum train weight which could be handled by two locomotives was thus 980 tons. From the increase in average train weights which occurred between 1906 and 1908, it may be deduced from table 10 that this fixed increment in capacity was being absorbed rapidly during these years. After 1910, the introduction of N-3's, each with a rating of 508 tons, increased the fixed increment in capacity obtained from the addition of one pusher to 508 tons, and raised the maximum train weight over the Selkirks to 1,016 tons. With the increase in average train weights recorded in table 10, it appears that this further fixed increment in capacity had been almost" completely absorbed by 1913, and that there was renewed pressure for a further increment in train capacity. Such a conclusion is reinforced by the fact that in the summer of 1913, that is, actually after the decision to build a double-track tunnel beneath Rogers Pass had been taken, but before the tunnel could be brought on stream, the C.P.R. undertook intensive dynamometer tests on the Mountain Subdivision in order to examine the possibilities for increasing train weights on the restricted sections. Figure 1 presents a profile of the alignment between Revelstoke and Beavermouth, and shows the 178 tonnage ratings which prevailed over the route both before and after the tests. 179 TABLE 10 AVERAGE TRAIN WEIGHTS, MOUNTAIN. SUBDIVISION, 1906-1913. (Equivalent Gross Tons) Year Westbound Eastbound 1906 539 668 1907 668 712 1908 743 787 1910 737* 1911 7901912-13 898 950 * Average for both directions. Sources:-1906-08: "Notebook," Kilpatrick MSS, Vancouver, n.p., n.d. 1910-11: Actual Gross Ton Miles Per Month/Total Train Miles. C.P.R. Co., "Statement of Train Locomotive and Actual Gross Ton Mileage, First District, British Columbia Division," Form S.O. 46, RCM. 1912-13: Cornell Civil Engineer, Vol. 23, No. 3, December 1914, p. 80. KEY :--1%—5»- Rising Gradient (1108) Tonnage Rating Before Tests 1016 Tonnage Rating After Tests W co EH O « O EH CO > 1% 1016 '(1108) EH S O pq S" t-H < < o 508 (508) 508 (508) > <; W pq > 2.2% FIGURE 1: PROFILE OF C.P.R. MAIN LINE BETWEEN REVELSTOKE AND BEAVERMOUTH, SHOWING TONNAGE RATINGS FOR SINGLE 210% LOCOMOTIVE BEFORE AND AFTER DYNAMOMETER TESTS, MAY 1913. (Ratings in Tons) Source:- F. J. Fisher to J. M. McKay, Superintendent, Revelstoke, May 25, 1913. RCM 76.15.188-45693.6 181 During the tests, which were held in summer conditions, "43 more tons was handled from Beavermouth to Rogers Pass at an average speed of 8 1/2 miles per hour,"6 but the "Engine slipped badly going through snowsheds."7 Eastbound, "35 more tons was handled Albert Canyon to Rogers Pass." However, since motive power was doubled on this latter section, as on the Beavermouth Rogers Pass section, the rating between Revelstoke and Albert Canyon had to be double the rating between Albert Canyon and Rogers Pass in order to ensure full utilisation of the single locomotive which would haul the train out of Revelstoke. This was found impossible to achieve, since the tests revealed that the existing rating between Revelstoke and Albert Canyon was already too high. The rating between Albert Canyon and Rogers Pass was therefore maintained at 508 equivalent gross tons, and the rating between Revelstoke and Albert Canyon was revised downwards to 1,016 tons.8 A train of 508 tons might comprise as few as ten loaded and two empty cars,' and even these ratings may have been reduced by five per cent, during the winter months.10 The tests demonstrated that the possibilities for increasing train weights over the summit of the Selkirks were extemely limited with existing motive power. Increments in train weight were offset by a severe penalty in train speeds, which at full throttle were already well below ten miles per hour for the heaviest freight trains. Moreover, the reduction in train speeds in turn curtailed the supply of train paths. However, since train weights could not be increased, the only way in which an increase in the total traffic volume could be accommodated was 182 by increasing the frequency of trains. This in turn increased the demand for train paths. b) Train Paths Estimation of the number of train paths which a single line of railway can supply is an imprecise science. The estimate is a complex function of transit time, which is in turn a function of motive power, train weight, train speed, opposing train movements and spacing of sidings; and of headways, which are in turn a function of the method of signalling, as well as of motive power, train weight and train speed. Computer simulation would be required in order to generate an accurate estimate of the capacity of the Rogers Pass line to supply train paths. Nevertheless, data are available for the demand for train paths in the years preceding abandonment of the surface route. Table 3 records the demand for train paths on both the Mountain and the Shuswap Sections during the years 1906-08. The period was one of declining traffic, and this table, with table 4, indicates that the decline in traffic was met by an increase in train weights, which in turn decreased the demand for train paths. Table 3 indicates that in 1906, the average demand for train paths was slightly over four per day in each direction. The table also indicates, however, that the demand for train paths through the Selkirks was highly seasonal, and more sharply peaked than demand on the Shuswap Section.11 In the summer peak, demand could reach almost six paths per day in each direction. By 1912, traffic volumes had increased to such an extent that the average demand for train paths was over eight per day 183 in each direction,12 and in the summer peak, the average demand may have reached almost eleven paths per day.13 Moreover, in 1912, the volume of traffic requiring transit through the mountains was forecast to double again within the next four years.1* An increase in traffic volume of such magnitude implied an increase in the demand for train paths to an average of sixteen per day in each direction,15 with seasonal peaks of almost twenty-two per day. It was because of these forecasts of future traffic volumes that the C.P.R. decided that its entire main line from Calgary to the West Coast would have to be doubled. Since other, less steeply graded portions of the main line were expected to be inadequate to handle this volume of traffic, it is reasonable to conclude that the alignment over Rogers Pass, with its forty-six miles of 2.2% gradient, was also deemed incapable of meeting the potential demand for train paths. From the available data on train weights and train paths, an approximate measure of the capacity of the main line may be calculated. There are three assumptions. First, it is assumed that the annual total of trains through Rogers Pass in 1912-13, the years upon which the C.P.R.'s evaluation of the Connaught Tunnel was based, represented the maximum number of train paths sustainable by the facility. Second, it is assumed that an increase in passenger traffic could have been accommodated by increasing only the weight, and not the frequency, of passenger trains: thus, an increase in passenger traffic would not decrease the number of train paths available for freight. Third, it is assumed that every freight train in either direction 184 conveyed the maximum tonnage operable by two locomotives, 1,016 tons. With these assumptions, it can be calculated that the maximum volume of freight which could have been conveyed by the 3,477 freight trains through Rogers Pass was 3,532,632 tons. In 1912-13, the C.P.R. actually conveyed 3,212,748 tons of freight through the Pass. This volume represented 91% of the calculated maximum tonnage which the facility could handle.1' This analysis demonstrates that the route over Rogers Pass imposed a potential bottleneck upon the flow of traffic through the mountains. This bottleneck was created by restrictions upon both train weights and the number of train paths available through the Pass. As traffic volumes increased, the surplus capacity which was available on individual trains was absorbed, until most trains were operated over the Selkirks at their maximum permissible weight, 1,016 tons. Since train weights could not be increased beyond this level, further increases in traffic volume had to be handled by increasing train frequency. The maximum number of train paths which the single line over Rogers Pass could supply appears to have averaged between eight and sixteen per day in each direction. The latter represents a non-sustainable average, since, by the time the demand for train paths reached this level, the C.P.R. expected that it would have to have double-tracked its main line through the mountains. The potential bottleneck appears to have been on the verge of becoming an actual bottleneck. This impression is reinforced by an analysis of traffic flows through Rogers Pass. 185 6.2 Traffic Flows This analysis will first consider total traffic levels through Rogers Pass, and then changes in specific traffic flows. Consideration of C.P.R. estimates of future demand will be deferred until section (6.6) of this chapter, although data concerning forecasted traffic levels are incorporated into the appropriate tables for purposes of comparison. a) Total Traffic Levels Table 11 presents the secular trends in the total tonnage of traffic requiring transit through Rogers Pass between 1904 and 1912, broadly disaggregated into freight and passenger volumes. Table 12 expresses these trends as annual rates of change. From the tables, it can be seen that passenger traffic, reacting less markedly to the recession of 1906-08, maintained a steady rate of increase throughout the years prior to the decision to abandon Rogers Pass. The volume of passenger traffic increased by 117% between 1908 and 1912. Over the same period, the volume of freight traffic, which had declined during the recession, increased by 123%. In gross tonnage terms, passenger traffic increased by 675,000 tons, while freight traffic increased by 2,575,000 tons. The tables indicate that in 1911-12, the rates of increase in traffic were the highest since 1908-09, the year of recovery from the recession. Moreover, in 1912, the year of the decision to double track through the mountains, the volumes of both passenger and freight traffic surpassed all previous records. 186 TABLE 11 GROSS TONNAGE OF PASSENGER AND FREIGHT TRAFFIC OVER EACH MILE OF ROAD, MOUNTAIN AND SHUSWAP SUBDIVISIONS, 1904-1913. (Equivalent Gross Tons) Year Passenger % Passenger Of Total 1904 0.325 21.7 1905 0.425 21.8 1906 0.475 17.0 1907 0.55 23.9 1908 0.575 27.1 1909 0.775 27.2 1910 0.875 26.5 1911 1.025 26.8 1912 1.25 26.6 1912-13* 2.002 30.9 1915-16** 4.314 32 Freight % Freight Total Of Total 1.175 78.3 1.5 1.525 78.2 1.95 2.325 83.0 2.8 1.75 76.1 2.3 1.55 72.9 2.125 2.075 72.8 2.85 2.425 73.5 3.3 2.8 73.2 3.825 3.45 73.4 4.7 4.471 69.1 6.473 9.155 68 13.469 * Average for two years, total tons over Rogers Pass. Cornell  Civil Engineer, Vol. 23, No. 3, December 1914, p. 80. ** Forecast. Sullivan to Bury, October 22, 1912, PIC CPCA. Source:- Histogram enclosed in, Bury to Shaughnessy, April 21, 1913, PIC CPCA (except 1912-13 and 1915-16). TABLE 12 ANNUAL RATES OF CHANGE IN GROSS TONNAGE PER MILE, MOUNTAIN AND SHUSWAP SUBDIVISIONS, 1904-1913. (% change) Year Passenger Freight Total 1904-05 + 30.7 + 29.8 + 30.0 1905-06 +11.8 + 52.5 +43.6 1906-07 +15.8 -24.7 -17.9 1907-08 +4.5 -11.4 -7.6 1908-09 + 34.8 + 33.9 + 34.1 1909-10 + 12.9 + 16.9 + 15.8 1910-11 + 17.1 + 15.5 +15.9 1911-12 + 22.0 + 23.2 + 22.9 1912-1912/13 + 60.2 29.6 + 37.7 Source:- Table 11. 188 Whilst freight traffic declined as a proportion of total volume, the absolute magnitude of the increase in freight tonnage, and the distribution of traffic between freight and passenger, which was heavily skewed towards the former, had serious implications for rail operations over Rogers Pass. An increase in passenger traffic might have been accommodated to some extent by adding cars to passenger consists and absorbing the surplus capacity which was available on passenger trains. By 1912, however, such an alternative was not available for freight operations. In 1912-13, every freight train which crossed the Selkirks required assistance. If traffic continued to increase, additional train paths and additional motive power would be required. In contrast, over one quarter of all passenger trains were still conducted by single locomotives, and could have borne some increase in tonnage without requiring an increase in paths.17 Table 13 presents the balance of traffic flows through Rogers Pass in those years between 1889 and 1913 for which disaggregated data are available. During the early years of operation, consistent with the C.P.R. forecasts made at construction time, westbound flows predominated. By the early years of the 20th Century, however, the balance had swung until the eastbound flows slightly outweighed the westbound. This pattern endured at least until the year of the decision to abandon Rogers Pass. Reversal of the C.P.R.'s forecasts would have implications for the C.P.R. system which are discussed below.18 189 TABLE 13 BALANCE OF TRAFFIC FLOWS THROUGH ROGERS PASS, 1889-1913. (Tons) Year Westbound  Tonnage % Of Total Eastbound  Tonnage % Of Total Total  Tonnage 1889* 38,895 1890* 50,773 1906 1,081,891 1907 1,100,610 1908 967,782 1910- 11** 1,710,779 1911- 12** 2,037,320 1912- 13 3,191,488 1915-16* 6,734,660 64 78 43 48 48 47 46 49 50 1,427,176 1,207,021 1,055,297 1,922,756 2,371,856 3,281,890 6,734,660 21,441 13,974 36 22 57 52 52 53 54 51 50 2,509,067 2,307,632 2,023,079 3,633,535 4,409,176 6,473,378 13,469,320 60,336 64,747 * Net tons of freight forwarded and received at Vancouver by rail, to and from the East. ** The equivalent gross tonnages of traffic given in Sullivan's letter to Bury have been disaggregated into passenger and freight volumes according to the' averages of the percentages of each for the years 1910 and 1911 and 1911 and 1912 respectively, according to table 11. The resulting passenger tonnages have been divided by 50 to yield estimates of the number of cars, and these estimates have been divided by 6.5 to yield estimates of the number of passenger trains. This was the basis upon which Sullivan himself projected future passenger traffic. The resulting freight tonnages have been divided by 508 to yield estimates of the number of freight trains. These estimates of the numbers of passenger trains and freight trains have been multiplied by 175 tons and 181 tons respectively, the weights of passenger locomotives and pushers (see Cornell Civil Engineer, op. cit., p. 80) in order to yield estimates of the weight of motive power which traversed the Pass. These estimates have been added to Sullivan's estimates of equivalent gross tonnages of traffic in order to yield estimates of total equivalent gross tonnages in each direction through Rogers Pass. # Forecast. Sources:- 1889-90: Vancouver Board of Trade, Annual Reports. 1906-08: Kilpatrick MSS, Vancouver, n.p., n.d. 1910-12 and 1915-16: Sullivan To Bury, October 22, 1912, PIC CPCA. 1912-13: Cornell Civil Engineer, Vol. 23, No. 3, December 1914, p. 80. 190 Within the Selkirks, partly because of the greater length of 2.2% gradient and curvature on the west slope than on the east, the time required in order to pass an eastbound train over the Mountain Section exceeded that required to pass a westbound train by some 20%.19 This may have aggravated the effect of the imbalance between eastbound and westbound traffic flows. The greater elapsed-time requirement for eastbound flows may have been symptomatic of actual congestion in the eastbound direction. However, there is no evidence to suggest that the existence of the imbalance between relative traffic volumes in itself caused the C.P.R. any concern. Neither does it appear that the C.P.R. expected any chronic imbalance in the future to exert pressure upon main-line- capacity in a single direction only. Although it is not known whether actual congestion was experienced on the main line through Rogers Pass prior to construction of the Connaught Tunnel, it is certain that the C.P.R. encountered some difficulty in handling existing traffic volumes. The Company's investment response to the 1910 avalanche disaster was motivated, not by an explicit desire to save human life in future, but by a recognition of the need to protect the expanding traffic. Thus, Bury had informed the Assistant Chief Engineer of Western Lines, F.F.Busteed: The way traffic is increasing on our main line, we must take steps to at least reduce delays, and it is necessary for you to prepare plans looking to minimizing trouble from slides.20 Bury echoed these sentiments to his President: "The increasing traffic makes it most necessary that interruptions be at least cut down to the minimum."21 It was not only during the slide 191 season, however, that difficulty was experienced: for almost two weeks during the summer peak of 1912, motive power had to be transferred from work-train to freight-train duty, and every locomotive on the Pacific Division, with the exception of four which were detailed to essential ballasting, was used to clear freight traffic.22 The nature and extent of the increases in total traffic levels through Rogers Pass may be more clearly understood by an analysis of changes in specific traffic flows. b) Changes In Specific Traffic Flows This analysis will examine changes in the following types of traffic through Rogers Pass:- passenger, lumber, grain, fish, other transcontinental traffic, and local traffic. (i) Passenger Few statistics are available concerning the actual number of passengers conveyed over the C.P.R. main line prior to construction of the Connaught Tunnel. Some impression of the magnitude of the increase in passenger volume which occurred during the period of surface operations through Rogers Pass may be gained from table 14, which contains the available data on passenger volumes for specific months between 1893 and 1908. 192 TABLE 14 PASSENGER VOLUME THROUGH ROGERS PASS, VARIOUS MONTHS, 1893-1908. Month Year Number Of Passenqers April 1893 4,609 April 1894 4,160 May 1895 2,549 May 1896 2,722 Oct. 1907 31,620 Nov. 1907 24,055 Dec. 1907 17,358 Oct. 1908 25,916 Nov. 1908 19,384 Dec. 1908 16,932 Sources:- 1893-94: Tait, Memorandum to Shaughnessy, May 8, 1894. PIC CPCA 1895-96: Tait, Memorandum to Shaughnessy, June 5, 1896. PIC CPCA 1907-08: "Notebook," Kilpatrick MSS, Vancouver, n.p., n .d. 193 It was the weight and frequency of passenger trains, rather than the actual number of passengers carried, however, which exercised the most important influence upon main-line capacity. Early transcontinental trains usually comprised a baggage-mail-express, coach, Colonist and sleeping car.23 By 1902, the only change to this consist was the addition of an observation car.24 In 1909, following gradient revision on the "Big Hill" and the' introduction of larger motive power, the first dining cars were attached to consists through the Rockies and the Selkirks,25 and by 1913, passenger trains were generally nine cars long,2' and averaged 443 tons in weight.27 Although a single road locomotive could haul a train of such weight, by 1913 three-quarters of all passenger trains were being pushed through the Selkirks in order to increase speeds and thus reduce line occupation.28 By 1916, however, the average speed of the fastest passenger service through the Selkirks, the eastbound train No. 2, was still less than twenty miles per hour,29 scarcely 1 1/2 miles per hour faster than it had been in 1902. Passenger train frequency also increased. During the summer of 1902, the single daily passenger service was augmented by an additional train in each direction on three days per week.30 In 1906, the double-daily service had to be introduced a month early because of increased demand,31 and it was maintained until October.32 During 1907, a third transcontinental service was added on three days per week,33 and this became a daily service during the summer of 1909.34 Until 1909-10, passenger service had always been pared to a single train daily between Calgary and Vancouver during the winter months. Although the C.P.R. had 194 intended to maintain a double-daily service throughout the winter of 1909-10,35 this was cut back from January to March 1910 between Vancouver and Calgary and between Winnipeg and Montreal.36 During the winter of 1912-13, however, the C.P.R. maintained its full complement of passenger services,37 and in the summer of 1913 increased the regular service to four transcontinental trains daily in each direction.38 It is likely that the number of passenger specials over the mountains, although unrecorded, also increased throughout the period, especially during the summer months. Thus, not only did the total number of passenger trains through Rogers Pass increase, but the seasonality of the scheduled traffic declined. Since passenger trains enjoyed priority over all other traffic on the main line, this increase in the number of passenger trains increased the probability that freight trains, also increasing in frequency, would have to be recessed in the limited siding accommodation, both ascending and descending the Pass, whilst awaiting "meets." Moreover, the probability was now extended into the winter months. Data for the transit times of "extra" freights is insufficient to permit analysis of the frequency and duration of recessing. Nevertheless, if the journey of an "average" train eastbound across the Mountain Section in 1908 required 12 hours 55 minutes,3' while the scheduled passenger trains made the crossing from Revelstoke to Donald in 6 hours 3 minutes, it is likely that the elapsed time for the journey of an "extra" freight would have exceeded 16 hours. Some of this prolongation in elapsed time may therefore have been attributable to the 195 creation of congestion as the volume of passenger traffic increased. (i i) Lumber Table 15 presents a record of the principal commodity flow through Rogers Pass by volume, that of lumber. Throughout the study period, the direction of the flow was almost exclusively eastbound, from the B.C. Coast and Interior mills to the Prairies, and, by 1909, as far east as the Great Lakes.40 The C.P.R.'s eastern lumber market began to develop in 1897-98,41 and provided an accurate index of crop conditions on the prairies. After the poor prairie crop of 1900-01 , 4 2 the years 1902 to 1907 were ones of marked expansion in eastbound lumber shipments by rail.43 By 1904, the trade was generating an average of 500-800 carloads per month, or at least one or two trainloads daily.44 The trade experienced recession in 1907-08,4s a fact which may explain much of the decline in freight volumes over the Mountain and Shuswap Sections during this period (see tables 11 and 12). However, the years 1910 to 1912 were marked by a dramatic increase in eastbound lumber shipments. This increase, unlike those of earlier years, which had been due to B.C.'s achievement of an increased market share in the prairies, was chiefly due to a large absolute increase in the volume of lumber consumed in the eastern provinces.46 Mirroring the trend in total freight traffic, the lumber flow peaked in 1912, the year of the decision to double-track the main line, and the rate of increase in the flow during the year preceding the decision surpassed all previous records. 196 TABLE 15 LUMBER TRAFFIC THROUGH ROGERS PASS, VARIOUS YEARS, 1900-1918. Year Total Board Feet Shipped By Rail From B.C. (Millions) 1900 26* 1906 360 1907 190* 1908 280** 1909 not available 1910 640 1911 761 1912 912 1913 756 1914 548 1915 562 1916 836 1917 537* 1918 530* * Rail shipments from B.C. Coast mills only. ** Annual output of Interior mills only. Source:- Vancouver Board of Trade, Annual Reports, except 1906. 1906: Victoria, B.C., Board of Trade, Annual  Report, 1906-07. 197 The lumber traffic was a key factor in the economics of the Rogers Pass route for four reasons. First, its sheer magnitude, in terms of both absolute volume and rate of increase, rendered it the most important of the trans-mountain freight flows, and a vital influence upon the capacity and the prosperity of the main line through Rogers Pass.47 Second, it was a distinctly seasonal flow, peaking in the summer months of constructional activity on the prairies, when main-line train paths through Rogers Pass were already at a premium on account of the summer peak in passenger traffic. Third, the loaded lumber movement was principally eastbound, against the greater length of 2.2% gradient. It was, therefore, more costly than westbound loaded movements, and imposed greater demands on main-line capacity since greater elapsed time was required in order to complete an eastbound crossing of the Selkirks. Finally, satisfaction of the demand for the loaded eastbound movement entailed a substantial movement of cars westbound. Since there was generally insufficient traffic with which to fill these cars, most were hauled empty, and as the demand for cars for the lumber trade increased, empties were hauled to B.C. from as far east as Fort William and beyond.48 Not only did the substantial, and expanding, empty westbound movement impose further pressure on main-line capacity through Rogers Pass, but it raised the joint cost of the total movement considerably. This increased cost was passed on to B.C. shippers in the form of higher rates, to an extent which goaded appeals from the shippers to the Board of Railway Commissioners on at least three occasions.49 198 (iii) Grain Grain shipments from the prairies westbound were negligible during the late 19th Century, due to crop failures50 and to the lack of markets accessible- via the western route. By 1903, however, the C.P.R. had shipped "several grain cargoes" to South Africa and Australia,51 and was competing to supply Japan and China.52 By 1909, markets had been developed in Mexico, the Philippines and Alaska,53 and the Tehuantepec Railway across Mexico was increasing in importance as a route by which grain shipped west from the prairies could penetrate the European market.5* The key stimulus to westbound grain shipments, however, was the westerly advance across the prairies of the grain production-frontier, which was extended beyond Medicine Hat in 1904.55 This afforded to westbound grain movements from the prairies to tidewater an increasing advantage over the hitherto predominant eastbound movements in terms of length-of-haul, car-cycle time, and therefore cost. By 1905, the Calgary Board of Trade calculated that, "grain from...Swift Current can be taken to Vancouver as cheap [sic] as to Fort William, probably for considerably less."56 The potential for westbound grain shipments, as a major commodity flow in themselves, and as a backhaul traffic which could offset the high joint costs of lumber movements through the mountains of B.C.,57 had early been foreseen. In 1902, the Edmonton Board of Trade had raised the question of Vancouver's suitability as a point of export for Alberta grain.58 Presentations canvassing the advantages of the western route were submitted to the Royal Commission on Transportation in 199 1905,5' and to the Royal Grain Commission in 1906,'° and both Commissions were convinced. The former concluded that, "a very material increase in westbound traffic from the whole province of Alberta may be expected.'&q