THE CRYSTALLOGRAPHY OF THE ROTATION SUBGROUPS OF COXETER GROUPS by NORMA BRODERICK B. S c , University of B r i t i s h Columbia, 1972 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n the department of MATHEMATICS We accept t h i s t h e s i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA A p r i l 1975 I n p r e s e n t i n g t h i s t Jaes is i n p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an a d v a n c e d degree a t the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e Head o f my Depar tment o r by h i s r e p r e s e n t a t i v e . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Depar tment o f M a t h e m a t i c s The U n i v e r s i t y - o f B r i t i s h C o l u m b i a V a n c o u v e r 8 , Canada ABSTRACT. The theory of space groups has i t s o r i g i n s i n crystallography and s o l i d state physics. In t h i s t h e s i s , we study those space groups i n n-dimensional Euclidean space whose point groups are the r o t a t i o n subgroups of c r y s t a l l o g r a p h i c Coxeter groups. TABLE OF CONTENTS CHAPTER I. 1. Introduction 1 2 . Some tech n i c a l r e s u l t s 5 3 . The crystallography of Coxeter groups 8 CHAPTER I I . 1 . A presentation of K + 12 2 . The case dim V K = 2 14 3 . The normal!ser of K + 15 4 . L a t t i c e s i n v a r i a n t under K. 18 5 . The groups H 1(K*,V/A) 2 2 6 . A p p l i c a t i o n to c l a s s i c a l crystallography 3 4 TABLE I 3 5 BIBLIOGRAPHY 3 7 ( i v ) ACKNOWLEDGMENT I would l i k e t o thank P r o f e s s o r s B . M o y l s and T . A n d e r s o n f o r r e a d i n g t h i s t h e s i s and P r o f e s s o r G . M a x w e l l f o r s u g g e s t i n g t h e t o p i c and h e l p f u l a d v i c e d u r i n g i t s p r e p a r a t i o n . 1 CHAPTER I 1. Introduction. The theory of space groups has i t s o r i g i n s i n c r y s t a l -lography and s o l i d state physics. A f t e r introducing the fundamental concepts, we s h a l l give a b r i e f h i s t o r i c a l o u t l i n e of i t s development. Let E be an a f f l n e space with n-dimensional r e a l vector space of t r a n s l a t i o n s V and a f f i n e group A(E). Once an o r i g i n has been chosen i n E, the elements of A(E) can be expressed i n the form ( t , g ) , where t e V and geGL(V). The composition rule i n A(E) i s then given by (t,g)(r,h) = (t+gr,gh). Let (x,y) be a p o s i t i v e d e f i n i t e symmetric b i l i n e a r form on V, P i t s orthogonal group and P(E) the subgroup of A(E) c o n s i s t i n g of isometries. A subgroup S of the 'Euclidean group' P (E) i s c a l l e d a space group i f S n v i s a l a t t i c e A i n V. The p r o j e c t i o n of S on P i s c a l l e d the point group K of S. Elements of K leave A i n v a r i a n t , f o r i f (a,g) i s a representative i n S of g € K and t € - A , the element ( a , g ) ( t , 1 K a . g ) " 1 = ( g t , 1 ) € S and therefore gt € A. We can thus regard S as an extension of A by K. unique cocycle K-—»V//\ can be associated with S. Let {(s(g),g)) be a system of representatives i n S of a l l g€K. If {(s'(g) ,g)] i s another such system, then (s(g),g)(s ,(g),g)'" 1 = (s(g)-s'(g ) , 1 ) €S, so that the function'B:K—^V / A obtained by reducing the values of s mod A i s uniquely determined by S. Since (s(g),g)(s(h),h) = (s(g)+gs(h),gh), we have 2 1s(gh) = s(g)+gs(h), so that's i s a cocycle. Conversely, given a cocycle " s : K — V / A > the set of elements i n T ( E ) of the form (s(g)+t,g), where g€K, t e A and s:K—=>V i s some l i f t i n g of "s, form the space group with point group K and l a t t i c e A which induces the given oocycle by the previous construction. Two space groups S and S' are considered equivalent i f they are conjugate by an element of A(E). I f S' i s conjugate to S by a t r a n s l a t i o n ( r , D , the.n.every element i n S' i s of the form ( r , 1 ) ( t , g ) ( r , l ) " 1 = (t+r-gr,g), where (t,g)£S. Thus'S» has the same point group and l a t t i c e as S but i t s cocycle "s'rK—^V / A d i f f e r s from s by the coboundary g — ^ r - g j * . Conversely, cohomologous cocycles K —•»V/A correspond to space groups which are conjugate by a t r a n s l a t i o n . Secondly, i f S' i s conjugate to S by an element i n A(E) of the form (0,h), then S' consists of elements (0,h)(t,g)(0,h)" 1 = ( h t , h g h - 1 ) , where ( t , g ) 6 S . I t follows that i n t h i s case A ' = h A , K« = hKh" 1 and s'(g) = h s U ^ g h ) . Consequently, the equivalence classes of space groups can be determined i n the following way. C a l l a subgroup of P c r y s t a l i o g r a p h i c i f i t leaves i n v a r i a n t a l a t t i c e i n V and determine the conjugacy classes of such subgroups i n GL(V). For each c l a s s {K] determine the set L(K) of a l l l a t t i c e s l e f t i n v a r i a n t by K and f i n d the o r b i t s i n t h i s s e t under the natural a c t i o n of the normaliser N(K) of K i n GL(V). Calculate H ^ K j V / A ) f o r each of these o r b i t s £ A ] . Let g.(K,A) be the subgroup of N(K) c o n s i s t i n g of those elements which leave A i n v a r i a n t . Then the o r b i t s of H 1(K,V/ A ) under the 3 action of H(K,_A) given by s (g) = hs(h gh) correspond to the equivalence classes of space groups with point group K and l a t t i c e .A. The determination of the equivalence classes of space groups dates back to the l a s t century f o r n ^ 3. In three dimensions, a space group can be considered as the symmetry group of an i n f i n i t e c r y s t a l and the problem of f i n d i n g a l l inequivalent space groups was attacked by crystallographers t r y i n g to understand possible c r y s t a l structures. In I 8 3 O , Hessel found the 32 classes of point groups; every c r y s t a l could now be placed i n t o one of 32 types according to i t s external symmetry. However, Hessel's work remained unnoticed for over 60 years and these groups were independently derived by Bravais i n 1 8 4 8 , who also found two years l a t e r the 73 inequivalent p a i r s (.A,K). In 1879 Sohncke determined the 5 4 inequivalent space groups whose point groups consisted only of r o t a t i o n s and observed that Jordan had derived them 10 years previously but had not expressed h i s r e s u l t s i n geometric terms. F i n a l l y , the complete l i s t of 219 inequivalent space groups was derived independently by Fedorov i n 1 8 8 5 , Schoenflies i n 1891 and Barlow i n 1 8 9 4 . However, i t was not u n t i l 1912 that Laue was able to confirm these p r e d i c t i o n s of c r y s t a l structure through experiments with X-ray d i f f r a c t i o n . In two) dimensions, a space group can be considered as the symmetry group of a plane pattern repeated p e r i o d i c a l l y i n two d i r e c t i o n s , such as i s common, for example, i n wallpaper and f l o o r t i l i n g designs. However, i t was not u n t i l 1891 4 that Fedorov gave the f i r s t mathematical treatment of these groups and determined the 17 p o s s i b l e inequivalent space groups. Fricke and K l e i n rediscovered these groups i n 1897, as did Po'lya and N i g g l i i n 1924. Examples of a l l 17 symmetry groups can be found among the decorative patterns of the early Egyptians. Indeed, i n Weyl's words, the a r t of ornamentation contains i n i m p l i c i t form the oldest piece of higher mathe-matics known to us. The s o l u t i o n of the problem f o r n ^ 3 l e d H i l b e r t to ask i n 1900 (as part of h i s 18th problem) whether there were only f i n i t e l y many inequivalent space groups i n any dimension. Fundamental r e s u l t s of Minkowski and Hermite implied that there were only f i n i t e l y many inequivalent p a i r s ( A , K ) . Joined to the f i n i t e n e s s of H U K .V/A), t h i s enabled Bieberbach and Frobenius to answer H i l b e r t ' s question i n the a f f i r m a t i v e . In 1951, Hurley [ l l ] published a l i s t of 222 point groups for n = 4 ; the l i s t was completed to 227 i h 1967. In 1965, Dade [9] found the maximal subgroups of GL^(Z) by considering automorphisms of p o s i t i v e d e f i n i t e i n t e g r a l quadratic forms. With the help of a computer, Bulow,fteubuser and Wondratschek [o"3 i n v e s t i g a t e d the subgroups of these groups and determined the 710 inequivalent p a i r s ( A ^ K ) . In 1972, Brown, Heubiiser and Zassenhaus found the 4733 inequivalent space groups. Ryskov [ l3] devised a method of constructing a f i n i t e family of p o s i t i v e d e f i n i t e quadratic forms i n n v a r i a b l e s such that a l l maximal f i n i t e subgroups of GL (.2) occurred among the 5 groups of t h e i r i n t e g r a l automorphisms, and a c t u a l l y found the maximal subgroups for n = 5« In the general n-dimensional case, Burkhardt [6] treated i n 1946 the s p e c i a l case of it being a symmetric group of degree n or i t s a l t e r n a t i n g subgroup, with A being a c e r t a i n type of l a t t i c e . Some progress was also made by Hermann fJO] ^ n 194-8* while Zassenhaus [15] developed an algorithm f o r the c a l c u l a t i o n of H^(K,V/ A ) . However, one cannot reasonably expect a general s o l u t i o n since every f i n i t e group w i l l eventually occur as the point group of some space group'. The c l a s s of point groups which are generated by r e f l e c t i o n s , 1. e. are c r y s t a l l o g r a p h i c Coxeter groups, has the advantage of being well understood and yet general enough to account (together with t h e i r subgroups) f o r a l l cases i f n ^ 3. i t was studied by Maxwell [jZ] . In t h i s t h e s i s , we s h a l l i n v e s t i g a t e the case when the point group i s the r o t a t i o n subgroup of a cr y s t a l l o g r a p h i c Coxeter group. Together with Maxwell's r e s u l t s , we can then derive p r a c t i c a l l y a l l inequivalent space groups for a ^ 3 from theorems v a l i d i n a l l dimensions. In the remainder of t h i s chapter, we s h a l l summarise some t e c h n i c a l r e s u l t s as well as the s i t u a t i o n when K i s a cr y s t a l l o g r a p h i c Coxeter group. The proofs of these f a c t s can be found i n [l2] . 2. Some t e c h n i c a l r e s u l t s . 2.1 I f K i s a subgroup of P , l e t be the subspace of elements f i x e d by K and V R the orthogonal complement of i n V.. The group K i s c a l l e d e s s e n t i a l i f s 0; i n p a r t i c u l a r , the r e s t r i c t i o n 6 K of K to V K i s e s s e n t i a l . Subgroups K and K' of T are conjugate i n r i f f they are conjugate i n GL(V); i t i s therefore s u f f i c i e n t to determine the conjugacy classes of point groups within T . Furthermore, K and K' are conjugate i n P i f f K and K' are conjugate by an isometry Vg—"^V.^,. As V s and V*K are i n v a r i a n t under H(K), we see- that \ •> N(K)"- =- ~ G L ( V K ) x N ( K ) . One can also prove that N(K) = C(K)N r, (K), where C(K) i s the c e n t r a l i s e r of K" i n GL(V^) and H.^ ( I C ) i s the normaliser of K i n fjj. 2.2 I f A i s i n v a r i a n t under K, we define the weight group J\ to be the subgroup of V con s i s t i n g of a l l elements x such that x-gx€A f o r every g^-K. The exact sequence 0—*A—vV-^V/A.—>0 gives r i s e to the long exact sequence ... —>H°(K, V) -5» H°(K, V/A) —>H 1 (K, A ) —>H 1 (K, V) —>... It follows that i f K i s e s s e n t i a l , we have A * / A = H 1(K,_A), so that _A / A i s f i n i t e and _/\ i s a l a t t i c e . A sub s t a n t i a l part of H*(K,V/_/Sj can be obtained as follows. C a l l a cocycle 1:K—>V/A. .weightlike i f a l l i t s values l i e i n _A*/A. The exact sequence 0—->.A /A—~>V/A — - > V/A _r> 0 gives r i s e to the long exact sequence 0 —>A**/ A -> Hom(K, A*/A )-^>H1 (KiV/ A ) —> The image of «^ w i l l be c a l l e d the group of weightlike classes and denoted by J L ( K , A ) . K 2.3 Suppose A i s a l a t t i c e i n v a r i a n t under K. Let A = A n V*, A K = A H V K and A Q a A K © A K . I f x = y+z € A, where y G V^ and z tV^., then x-gx = z-gz £ AR% s o that z<?A*. The mapping x —>z induces a homomorphism A/A~—~?A*/AV which i s i n j e c t i v e since i f z , y = x-z £_AK and thus x <fA_ 7 I t follows that A / A i s f i n i t e and A i s a l a t t i c e . Let ©(A) be the image of A / A 0 i n -^/•Ar T h e m a P P i r i S x—>y also induces an i n f e c t i v e homomorphism J^/A Q -*> V ^ / y \ K whose image i s isomorphic to ©<A>. Conversely, given l a t t i c e s J\K C and -AR C V K and a subgroup ® of A g / A K which i s isomorphic to a subgroup ®^ of V^/A^» one can construct a l a t t i c e _A ^>A0 = A K @ A £ such that (9(A) = 0. Namely, l e t K : 0—* 0^ be some isomorphism and define where z Q ( r e sP»y^( Q)) a r e representatives of the elements of ©(resp. (H) ) i n V £ (resp.V^). One can show that i f A* i s another l a t t i c e i n v a r i a n t under K, then A* =gA f o r some g£N(K) i f f there e x i s t s h e N(K) which maps -A f i to A£ and induces an isomorphism @(A) —*> Q ( /V ) • Therefore, i n order to obtain a set of representatives of the o r b i t s of L(K) under the acti o n of N(K), i t i s s u f f i c i e n t to f i r s t choose a set of l a t t i c e s A^ representative of the o r b i t s of L(K) under K(K) and an a r b i t r a r y l a t t i c e A K C V^. For each A K > one then chooses a representative set of subgroups © of A* /Aj^ from those which are isomorphic to subgroups of F i n a l l y , one combines j\ , a n d & by (1). 2.4 To c a l c u l a t e H^K.V /A) , we s h a l l e s s e n t i a l l y use the idea of Zassenhaus Cl53» Choose a presentation F/R of K with generators fe^i^I 3 1 1 ( 1 r e l a t i o n s L ^ j l j ^ j * B v f i r s t reducing the elements of F modulo R, one obtains an acti o n of F on V. A cocycle t:F—=>V i s uniquely determined by the vector ( t ( e i ) ) , which may be chosen a r b i t r a r i l y . Since t(uR.u _ 1) » ut(R.j, the values of t on R are determined by the vector ( t ( R . ) ) . 0 8 The cocycle t : F — ^ V induces a cocycle "t:K —>V/A i f f t(R.) £ A J for a l l j € J , and every cocycle K — ? V / A can be obtained i n t h i s way. 2.5 I f the i n v e r s i o n - 1 ^ i s not an element of K, we s h a l l denote by +K the subgroup generated by K and - l y C l e a r l y L(K) = L(+K) and N(K) C N(+K); however, there may be l a t t i c e s inequivalent under N(K) but equivalent under N(+K). I f A €L(K), i t follows from the exact sequence ° ^ H ^ + 1 V , ( V / A ) K ) ^ H ^ K . V / A ) - ^ H H K . V / A ^ V that H ^ ( + K , Y / A ) i s isomorphic to the d i r e c t sum of the group of elements of order ^2 i n H*(K,V/A) and the quotient of A * / A by 2 ( A * / A ) . 3, The crystallography of Coxeter groups. Suppose K i s the Weyl group W(R) of a root system R i n V K. Following Bourbaki L 1 3, we denote by s ^ the r e f l e c t i o n corresponding to the root <*. and by Q(R) and P(R) the root and weight l a t t i c e s of R. Let B = be a b a s i s of R and S = [ s i ] the corresponding set of generators of K, where we extend the function from to V by l e t t i n g i t vanish on V*. Then {(s±s^) ^ = 1] i s a presentation of K, summarised by the Coxeter matrix ( m ^ ) . 3.1 We define X(R) to be the group of those g€GL(V^\ which s a t i s f y goi = a ^ hoi f o r a l l o(f B, where h i s an angle preserving permutation of B and the are p o s i t i v e r e a l numbers s a t i s -f y i n g n( p ,c« ) a Q L = n(h(3 ,hc< f o r a l l c B. Then N(K) i s equal to the semidirect product of *K and'A"'(R). I f R i s i r r e d u c i b l e , one can see that 1(R) = H.^(R), where H i s the group of p o s i t i v e homotheties and AQ(R) the group of graph automorphisms of B except f o r the cases of Bg'^S ^ » when AQ(R) posseses an a d d i t i o n a l element. 3.2 Every l a t t i c e -A^ £ L(K) i s such that Q(R') P(R«) and A K = P(R') f o r some root system R' s a t i s f y i n g W(R') = K. I f R and R* are two root systems of the same type for which W(R) = W(R«) = K, i t follows from 3.1 that there e x i s t s g€A(R) such that R' = gR. Therefore l a t t i c e s s a t i s f y i n g Q(.fi') ^ A £ C P(R') are equivalent under N(K) to l a t t i c e s A £ such that Q(R) C A £ ci P(R). Consequently, i n order to f i n d a representative set of l a t t i c e s , i t i s s u f f i c i e n t to consider only d i f f e r e n t types of root systems R f o r which W(R) = K and f o r each type R only those l a t t i c e s A s a t i s -K fylng Q(R) C AK C P(R) and -A* = P(R); the l a t t e r condition i s equivalent to r e q u i r i n g that U/2. f A for a l l <* C B. 3.3 By 2.4, a map t:S—->V w i l l induce a cocycle~E:K —*>V/A i f f <2) t U s ^ ) ± V = ( l + s ± s >...+( s ^ ) ^ ) ( t ( s i ) + 8 j L t ( g i ) ) € f o r a l l i , j . Let U s ^ = y ± + ^ k t k i < * k , with y ^ V 1 ^ and define p.^ = < t ( g i ) , . Then t i s a coboundary i f f t ( s . ) = t . . o i . mod A f o r a l l i . One can deduce from (2) X\ X X X. that 2.H 1(K,V/A) = 0. By subtracting from t the map b( s i ) = P-y^ ^j,/2, which induces a coboundary, we can assume that p. . € 2 • Equations (2jl then amount to the following: (3) 2 t ( s i ) € A f o r a l l i (4) p ^ o r f j = p j . ^ i mod A 0 (5) t ( s ) - t ( s . ) = p.. ex. _ p ^ m o d A i f n. . = n.. = 1 0 31 0 i j 1 xj jx 10 (We have abbreviated the Cartan integer n( tf., oi .) by n. ..) X J X j Since the projection of A on Vg. i s contained i n •AK = P(R), we can evaluate **. at both sides of eqns (3)-(5) to obtain u the following necessary conditions for t to induce a cocycle: (6) 2p±i £ H for a l l i , j ( ? ) P i j n j k " * j i n i k ^ ^ k " ^ i j » 0 ( 8 ) Pik-^jk = 5 j i : n 3 k ~ p i j n i k m o d 1 f o r a 1 1 k i f n i j = n j i = Call the last root c<m i n a conjpnent of B of type C exceptional: i t i s characterised by the fact that 0<m/2 i s a weight. The equations (6)-(8) imply that by r e p l a c i n g " with a cohomologous cocycle, we can assume that P^€ 2 except for the following cases (when i / j ) : (i ) cx! and ot. are exceptional and ( . - ot .)/2 4 A , when p i j = p 5 i m G d Z ' ( i i ) o( i s the last and oi. the second last root i n a component X J of type C , m ^ 2, when ^ A A ) = 2L, ( i i i ) ot^ and cx1.. belong to a component of type A^,B^ or B^ and ( ot^- 0^)./2 6 j \ , when the only exceptions are p ^ = p^j = p ^ mod (iv) c*!.. and o( belong to a component of type D. , when the exceptions can be p ^ = p 3 1 = p 2 3 mod ^ , P l Z f = P2fl = P 2 Z f mod £ or p ^ = p ^ = P 2^ + P 2^ fflod- Z • Correspondingly, the elements (cx^- o<5)/2, ( o^- «^ )/2 or (tf-j- ^)/2 must belong to A . 3.4 When p. . 6 Z for a l l i , j , the cocycle *t i s weightlike i n the sense of 2.2. The dimension over W/Z% of the subgroup «TL(K,.A) can be calculated as follows. Let & be the dimension over W/2. Z of the set of elements of order ^ 2 i n ( A P ( E ) ) / A , P the number of connected components i n the Dynkin diagram of B after 11 a l l double and t r i p l e bonds have been removed and T the number of components of type C m i n B. Then dim^-H-CK,A) = <£p - X . 3 . 5 I f TT i s a component i n B of type C^, with m odd and 3 , then <V m_.j£Q(R) so that condition ( i i ) of 3 . 3 i s s a t i s f i e d . Let t n:S-*>V be defined by t ( 6 f f i ) = <•> /% when co^ i s the l a s t root of n and t ( s ^ ) = 0 otherwise,; then t induces a cocycle "t :K—->V//V . Let "C be the subgroup of H 1(K,V /A) generated by the cohomology classes of such cocycles. Then ~t has zero i n t e r s e c t i o n with J 1(K , _ A ) and i t s dimension over 7?/2.H i s equal to the number of components i n B of type C^, with m odd and 3 » Frequently, JL(KtA) <£>C accounts f o r the en t i r e group H 1(K,V /A) . More generally, suppose £R q] i s the set of components i n B of type U-^ , A^,B^,B^ or and R' the remaining part of R. Suppose that A = - A K @ A ^ @ © q - A g ) , w h e r e ( r e s p . A ^ ) i s the i n t e r s e c t i o n of -A^ with the subspace spanned by the elements of R« (resp.R q;. Then n J U j V / A ) = J1(K , A ) @ £ G>B where the dimension of £) over 2 7 2 % i s equal to the number of components fiq of type A ^ » B ^ Q r B^ plus twice the number of components of type for which A^ equals P ( R q ) . For l a t t i c e s not covered by these remarks, the existence of a d d i t i o n a l cocycles can be inve s t i g a t e d i n each case by s t a r t i n g from the equations ( i ) - ( i i i ) of 3 . 3 . 12 CHAPTER I I Let K be a c r y s t a l l o g r a p h i c subgroup of P generated by r e f l e c t i o n s and K + i t s r o t a t i o n subgroup. In t h i s chapter, we s h a l l i n v e s t i g a t e the crystallography of K +. Suppose R i s a root system i n V R for which W(R) = it. Let B = be a basis of R, S = fs^] the set of corresponding generators of K and ML = C^-j) the Coxeter matrix of S. We s h a l l abbreviate the Cartan integer n< o(^, d ) by n ^ and extend the function of^ from V^ . to V by K l e t t i n g i t vanish on V . The case of dim = 1 i s t r i v i a l and w i l l be excluded from discussion. The somewhat a t y p i c a l case of dim = 2 w i l l be discussed i n section 2 ; from then on, we s h a l l assume that dim "7/ 3 » 1 . A presentation of K +. The group K + consists of those elements of K which can be written as a product of an even number of elements of S. Choose any cV ^ £ B and l e t g^ = s i s o f o r ^ °» t l l e n S^€^*» Since g^S^ = 6 i s j * t n e s i 6 g e r i e r a t e K + * They s a t i s f y the following r e l a t i o n s : CD g ^ 1 0 = 1 , iSi&Jy1* = i . 1 .1 Proposition. The r e l a t i o n s ( 1 ) form a presentation of K.+ . Proof. Let G + be the abstract group defined by a set of generators {g^J » subject to the r e l a t i o n s (l). Then fcCg^ = g*[ + 2 defines an automorphism of G such that £ = 1, since a word of the form gj~ i s mapped to i t s inverse by £ , while a word of the form (g.g't 1 ) m i j i s mapped to ^ Ug^g^' i^^g"^ We, can obtain an action of the two-element group £+lJ on G + cby - 1 13 a s s o c i a t i n g t h e automorphism £ t o t h e e lement - 1 . L e t G be t h e s e m i d i r e c t p r o d u c t o f G + and[+l] r e l a t i v e t o t h i s a c t i o n . The f u n c t i o n ^(g^) = ^ s 0 and ^C -1 ) = s Q i n d u c e s a homo-morphism G —>K s i n c e t h e c o n j u g a t e o f by -1 i n G i s mapped t o the same e lement a s fCg^) = g^> namely s Q s ^ . C o n v e r s e l y , t h e f u n c t i o n ^ ( s ^ ) = ( g i , - l ) , f o r 1 / 0 , and ' / ' (SQ ) = ( 1 , - 1 ) i n d u c e s a homomorphiism K — ^ G s i n c e an e l e m e n t o f t h e form 4 s i s 3 ) ' * o r ^ 0 i s m a P P e d t o ( ( g I » - D ( g ; j » - 1 ) ) = ( ( g ^ ] 1 ) ^ 1 ) , w h i l e ( S ^ S Q ) 1 0 , f o r i / 0 , i s mapped t o m i 0 m i 0 ( ( g ^ ' - l K l j - l ) ) = (g^ » 1 ) « A s ^ i s c l e a r l y t h e i n v e r s e o f j& , t h e l a t t e r must be an i s o m o r p h i s m . The r e s t r i c t i o n o f 0 t o G + i s t h e r e f o r e an i s o m o r p h i s m between G + and K + . / / The r e l a t i o n s {1) a r e w r i t t e n down by C o x e t e r and M o s e r [ 8 > P » 1 2 4 ] who do n o t , h o w e v e r , make any comment a s t o whether t h e y c o n s t i t u t e a p r e s e n t a i o n . The above method o f p r o o f i s s u g g e s t e d by an e x e r c i s e o f B o u r b a k i f I ,p .38 J. 1.2 P r o p o s i t i o n , I f K i s n o t o f t y p e B 2, t h e n K + / [ K + , K + ] (Z^Z)?" 1 © (273Z)7" , where p i s t h e number o f c o n n e c t e d components i n t h e D y n k i n d i a g r a m o f B a f t e r a l l d o u b l e and t r i p l e bonds have been d e l e t e d and = 0 , u n l e s s K i s o f t y p e A ^ G ^ A ^ o r D ^ , when 1. I f K i s o f t y p e B 2 > the g r o u p K + / f K + , K + J = Z ' A Z . P r o o f . We o b t a i n a p r e s e n t a t i o n f o r K + / ( K + , K + ] by a d d i n g the r e l a t i o n s g . g . = g . g . t o ( 1 ) . Choose o^-. t o be c o n n e c t e d t o a t most one o t h e r r o o t c B . I f °<j, / ^o* °^1 t h e n s i = 1 ' — 1 2 i f t h e r e e x i s t s c^^ d i s c o n n e c t e d f rom oi^t t h e n (g^g^ ) = 2 , so t h a t a l s o g ^ = 1. The r e l a t i o n s ( g 1 g ' ! 1 ) l t t i j = 1 t h e n amount I X J t o s a y i n g t h a t g ± = g_. i f m a 3 and a l s o t h a t a 1 i f « a 3j the conclusion follows. In the remaining cases, i f dim = 2, K i s of type ^ 2 , B 2 or and K + i s correspondingly c y c l i c of order 3,if or 6. I f K i s of type B^, then g^ = 1, g 2 = 1 and ( S ^ g ^ 1 ) ^ = U whence g^ = 1. I f K i s of type A^ or D^, then g^ = i , g | = ,l,and ( g ~ 1 g ± ) 3 = 1 for i > 1 , so that g ± = 1.// We s h a l l often need to use the formula (2) s ^ v x ) = x - <x, - i ^ * ^ - n d i<x, c*\ . I t follows for example that V* = v* since s.s.(x) = x for a l l c/. ^ cx\ i n B implies <x, ©(A = 0 for a l l B, so that X £ V*. 2. The case dim V K = 2. By 1.2.1, H(K +) = GL(V K)N(K +), where K + i s the r e s t r i c t i o n of K + to Y^; i t therefore s u f f i c e s to determine H(K+) i n case Y s: Yg.. 2.4 Proposition. I f K i s e s s e n t i a l . N(K +) i s the group of s i m i l a r i t i e s of V i f R i s of type A 2,B 2 or G 2 and a l l of GL(Y) i f U i s of type A j X A ^ Proof. I f R i s of type A^XA^, the group. K* consists of|/+1^ so that H(K +) = GL(V). Otherwise, K + i s generated by a r o t a t i o n g such that g 2 / 1. By 1.2.1, we have N(K +) s C ( K + ) N p ( K + ) . The group H p ( K + ) contains the e n t i r e group of r o t a t i o n s P+ since the l a t t e r i s abelian. I f s i s a r e f l e c t i o n of Y> so i s gs and therefore ( g s ) 2 = 1 or sgs = g""1£ K +. Thus B p ( K + ) = P . In general, i f f = pu i s the polar decomposition of an a r b i t r a r y f £N(K) with p hermitian and u €. P, we have p £C(K +> and H ^ ( K + ) . Since the minimum polynomial of g i s of maximum possible degree, C(K +) i s spanned by 1 y and g. Suppose p = a1 y+bg for some a,b€lR ; then the adjoint p* = al^+bg"' = p, which i s 15 only possible i f b = 0 , singe ^ £ 1. Therefore f i s a s i m i l a r i t y of V.// When V, = V K, an elementary argument (e.g. [6,p.55^] ) shows that ( i ) i f R i s of type A^XA^, every l a t t i c e i s invariant under K + ( i i ) i f R i s of type A 2 or Q^t every hexagonal l a t t i c e i s i n v a r i a n t and ( i i i ) i f R i s of type every square l a t t i c e i s i n v a r i a n t . In each case, prop.2.1 shows that every l a t t i c e i s equivalent to the root l a t t i c e Q(R). From eqn . ( 2 ) , one sees # that Q(R) = P(R). When V K £ 0 , the determination of invariant l a t t i c e s proceeds according to the general p r i n c i p l e s of I. 2 . 3 * I f -A. i s a l a t t i c e invariant under K +, the group H 1(K +,V /A ) can be calculated by the general method available for c y c l i c groups. Let g be a generator of K +, of order m. Since 1+g+...+g ~ vanishes on V K, while the r e s t r i c t i o n of 1-g to V K i s i n v e r t i b l e , one sees that H 1(K +,V / A ) = { V e ^ M * I my = o] / ®^A) . 3 . The normaliser of K +. We s h a l l prove i n t h i s section that H(K +) = W(K) i f dim V K > 3 . Since N<K+) = GL(V K)H(K +) and H(K) = GL(V K)H(K), i t i s s u f f i c i e n t to prove the assertion i n case V = Vg. 3 . 1 Proposition. C(K +) = C(K). Proof. I f f € C(K +), then f s s (x) = s s f(x) for a l l i n B and a l l x£V. Using ( 2 ) , t h i s can be written as <x, «p Uoi.) + «x, Z^y- n ^ x , ^ d»f((X i) = <f(x), ^ > o ( . + (4(x),o( i> - ^ . ^ x ) , ^ ) ^ . I f we choose x to s a t i s f y <x, ot ^ = 0 and <x, = 1, we see 16 that f( o(^) belongs to the plane P^ .. spanned by crt^ and c**^ .. I f / ^ » t i l e same argument shows that ^ ^ i ^ C P ^ ; conse-quently f(«* ..) €B±2 H P i k a [RoCj,. Suppose f ( o^) = a ± cy ± for some a^e (R. Letting x = o/^ i n the above equation and comparing the coefficients of ot ., we see that n. .a. = n. .a.. It follows U X J j J X that a^ •= a., whenever and belong to the same component of B. Therefore f€C(K).// Since N(K +) = C(K +)Up(K +) and N(K) = C(K)Np<K) by 1.2.1, i t remains to prove that H p (K ) = Np(K), We need the following well-known result (e.g. [7,p.283): 3 .2 Proposition. Suppose w £ K and l e t V w = (vc V I w(v) = v j . Then w can be expressed as a product of reflections corresponding to roots i n the orthogonal complement of ¥ w i n V..// For each pair of independent roots cK, R, l e t be the plane spanned by o(. and p and l e t P be the collection of a l l such planes i n V. 3 . 3 Proposition. An element g £ N p ( K + ) permutes the elements of P. Proof. Suppose ol,p € R, with c* / + p , and Consider the element w = gs^ Sp g~ 1£ K+. Since g € P , the subspace V w = g(H^ r> Hp ) i s of codimension 2 i n V and i t s orthogonal complement i n V i s the plane U = gCP^^ ). Prop.3.2 implies that w = 6 ^ Sj- ... for some roots If , J ,... i n U. There are at least two independent roots among the , since otherwise w, being i n K+, would equal 1. Therefore U = P^^ for some roots € U . / / 3 . 4 Proposition. Suppose and F ^ are distinct elements of P such that the subspace L = P^ + P^j- i s three-dimensional. Let R' be the root system R HL. Suppose that either (a) R' i s reducible or (b) there exists a rotation w i n P of order ^ 17 and a r o t a t i o n w2 i n P y j of order 3 or 4 such that ( W j V ^ ) = 1. Then P ^ O P ^ j i s spanned by a root p € R. Proof. I f R' i s reducible, i t i s easy to see that every i n t e r s e c t i o n P^p n Py^ i s spanned by a root. On the other hand, a well-known formula of s p h e r i c a l trigonometry asserts that i f r^ and r 2 are rota t i o n s of L through 0^ and © 2 , and i f t h e i r planes are i n c l i n e d at an angle ft , then the angle © of the r o t a t i o n i s given by cos 0/2 = -cos ©^2 cos © 2 /2 + cos f£ s i n Q^/2 s i n © 2 /2 . Applying t h i s formula to the rot a t i o n s w1 and w2, we conclude that the angle <£> between P^ and P ^ i s given by cos f6 = 1/3 or \/\f5* An e x p l i c i t look at root systems of type A ^ B ^ and now reveals that t h i s i s only possible i f ^ o Py^ i s spanned by a root.// 3.5 Proposition. H p ( K + ) = H p ( K ) . Proof. We have seen i n 3*3 that an element g e H p ( K + ) permutes the elements of P. I f B has no components of length ^3, we choose any three elements o^, <*2, oi^CB and consider the planes P^ p = g P 1 2 and PJ-J- = g P j y Since P 1 2 + P ^ i s three-dimensional and R' ne c e s s a r i l y reducible, prop.3.4 shows that F ^ ^ n i s spanned by a root p . Since ? 1 2 n p i s spanned by ^, we conclude that g maps ^ i n t o a multiple of p and therefore g s ^ " 1 = Sp € K. As g s ^ ^ ' U K * f o r a l l i , we have g s ^ " 1 ^ K, so that g belongs to N p ( K ) . I f B. has a component of length >3, then we can choose the elements c*^ ( i = 1,2,3) i n such a jray that m 1 2 = 3 and m^ = 3 or if. I f we l e t w] = g s ^ g - 1 and W2 = ] S S i s 3 S " ' a n d a P P l y propO'Jf, the argument proceeds as before.// 18 4. L a t t i c e s i n v a r i a n t under JS. . We s h a l l assume that V = V R j the general case can be treated by the general method o i I.2.3* I t turns out that a l a t t i c e i n v a r i a n t under K + i s also i n v a r i a n t under K, apart from the following exceptional cases; (a) Suppose R i s of type Aj X... XA^. Let A 0 be the s u b l a t t i c e of P(R) containing Q(R) and a l l elements of the form ^j ) /2 f o r c<^ y o<^ i n B. When dim V i s even, define Al = A e U < A e + ( %1 A* • A U ( A e + ( ZkcXj.JA • V2) • (b) Suppose R i s the d i r e c t sum of two root systems R^ and R., of types A 1 and A 2 r e s p e c t i v e l y . Let A Q = QC.Rj ) $ P( Rg) and define A l = A e U <^e + .< V °y/3> U ( A " < °S + °y^> - A e 2 = A e u ( A e + - <x2)/3) tl ( A e - (<x r cx2)/3) , where {c< ^ and ^O1^, °*3$ 3 1 , 6 bases of R1 and Rg. (Crystallograph c a l l these l a t t i c e s •rhombohedric'.) (c) Suppose R i s the d i r e c t sum of two root systems R^ and 82* both of type A_. Let A = P(R) and define 2 e _ A e l s A e u ( A e • ( o y o y / 3 ) U ( A g . ( o t i + ^ 3 > / 3 ) A e 2 = A s ^ ( A Q + ( c < r ^ ) / 3 ) u ( A e - t c y r (X 3 ) /3), where {c^j» o i 2l a J l d ^*3* °Sf^ a r e bases of and R 2« In each of the above cases, the l a t t i c e s A., and A . " el e2 w i l l be c a l l e d o l exceptional type w.r.t, R. One can v e r i f y that + they are i n v a r i a n t under K , but interchanged by an element of H not i n K ( and thus equivalent under fl(K ) ) . 19 Suppose / i i s a l a t t i c e i n V. i n v a r i a n t u n d e r K + . F o r e a c h oi±€ B , l e t -A = A O and d e f i n e = ® ±S^. R e c a l l t h a t A* = ( v C V | v - gv<?A f o r a l l g ^ K + j i s a l s o a l a t t i c e i n V w h i c h c o n t a i n s A . 4.1 P r o p o s j L t i o n . Suppose o ^ , and c < k a r e d i s t i n c t e l e m e n t s o f B s u c h t h a t a... = 0. Then i f x € J\ , we have — — — — — J K (a ) i f n ^ = n ^ = 0, t h e n 2 <x, ^7 <X± ^ -A. (b) i f n 5 J L = -1, t h e n ^ - a ^ n ^ T a ^ n ^ ) < x , e S ^ o ^ A . A * P r o o f . S i n c e x € A , t h e e l e m e n t s x - s ^ C x ) = <x, <&k> * k + « x , oJ ±> - n ^ x , ^ j ^ ) ^ s^SfcCx) - x = -<x, &..> - <x, <*k>0<k a l l b e l o n g t o A . C o n s e q u e n t l y t h e i r sum (2 <x, c ^ > - n J ± < x , ^ > - n ^ <^x, £ k » « ± a l s o b e l o n g s t o A . I f a.. = a. . = 0, we o b t a i n { a ) . O t h e r w i s e , we J x K l a p p l y t h i s c o n c l u s i o n t o t h e e lement y = ( x - s i s j ( x ) ) - ( x - s ; j s : L ( x ) > = - n ^ ^ x , ^ • ^ ) ° L ± + n i j ° ^ i n s t e a d o f x and deduce ( b ) . / / Note t h a t t h e i n t e g e r 4-B.j j n j j , " * n j k n k j i n P a r t c a n o f l l y assume t h e v a l u e s 1,2 o r 3. I f A- i s o f e x c e p t i o n a l t y p e w . r . t . R, one c a n v e r i f y t h a t A # = Q ( R ) . it * 4.2 P r o p o s i t i o n . Suppose A = Q(R) and x <£ A i s s u c h t h a t </x, 4- Z f o r some o(^$B. ( a ) I f 2 <<x, ot^ £ Z , t h e n R i s o f ty,pe A^x ... and e i t h e r _A = -A o r d i m V i s e v e n and A e q u a l s A , o r A ~ . F u r t h e r m o r e . e ~ ~ : e 1 — — e2 * x = ( X ± o < i ) / 4 mod P ( R ) . 20 <b) I£ 3 <x, € Z , then e i t h e r ( i ) R i s of type y Ag, -A = -A e 1 or A e 2 and +x = ( « t + c<2)/3 mod P(R) or ( i l ) R i s of type AgXAg, A = AQ^ or A e 2 and +x = ( 0 ^ + ^) /3 mod P(R). Proof. For any ^ i n B, consider (5) x-s is^(x) s <x, <^j> °^ + (^x, <*±> - n^^x, e^)) ^ 6 A . I f H<x, 2T for some o<,.€B and some integer N, m u l t i p l y i n g (3) by N shows, i n view of A # = Q(R)» that N.<k, o f ^ C Z f o r a l l c*^<B. Thus the hypothesis of the proposition implies that <x, ot. /> 4- 2 for a l l <x\ CB. I t now follows from (3) that V (#) <x, ei ±> - <x, ^ > I Z, since otherwise <x, ©ci.j> would have to be i n Z . v _ Suppose we are i n case (a); then 2<x, c t . ^ Z f o r a l l i-> o ^ C B . Eqn.(J+) implies that n ^ i s always even, which can only happen i f R i s of type A^X ... XA^; furthermore, (3) shows that AeC A. I f y £(A HP(R))\Ae, then y = C XQ / 2 mod A g, contrary to A # = Q(R), so that AnP(R) = A . Since <x, = i mod Z for a l l oi. € B, we have x = ( Z . ) A m o d P(R). I f dim V i s odd, x cannot belong to A- since then 2x = s<g/2 mod A would also belong to A. As R i s of type A. * ... *A.,, prop.if. l(a) shows thatT^2<y, € Z . I f dim V i s odd, i t follows that A. CP(R) so that A = _A. I f dim V i s even and A <f- P(R), then A. must be one of the l a t t i c e s A e 1 and - A e 2 since ( ^ o ^ ) A and ( 5 ^ . 0 ^ ) A + c*Q/2 cannot both be i n A . In case (b), we have 3 <x, C Z for a l l c / ^ B . Eqn.(if) implies that </x, tL^y = <x, t^^"^ mod Z whenever n ^ = - 1 . Further-more, 4.1 shows that ( i ) B does not contain three disconnected roots and ( i i ) B does not contain roots connected by double or t r i p l e bonds or a sequence of three roots connected by s i n g l e 21 bonds, since otherwise <x, o^/ or 2 <x, would belong to "JL for some 6 B. This means that R i s ei t h e r of type A, xA, 1 1 2 or A-*A-. In both of these cases, (3) implies that A--^ -A . c. c. e Suppose R i s of type A^ x, k^. Then x = x^ ^ + x 2 ( < 2^+ ^ ) mod P(R), where x p = +1/3 for p = 1,2 and ^ i s the weight corresponding to © C ^ . Since 6 ^ / 3 = = - ^ / 3 mod P(Rj) and (A^ 2+ ^ ) / 3 = ( <*2+ <*3)/3 = - mod PCR^we have +x = ( cx^i oi 2 ) / 3 mod P(R). Eqn.(3) applied with i = 1 and j = 2 shows that ( 0 ^ + C ^ 2 ) / 3 £ A with the same choice of sign as f o r the value of +x mod P(R). Both of these elements cannot belong to A since t h e i r sum 2 0^/3 ^A. I t follows that e i t h e r -A e |CA or A e 2C_A. I f y £ ( A n P ( R ) ) \ A then y = c^ / 2 mod A contrary to A # = Q(R). Therefore A n P(R) = A . I f ycA\p(R)> choosing i = 2, j" = 3 and k = 1 i n prop.if. 1(b) shows that 3 <y, <* N £ 2T. I t follows that A , = A or A , = -A.// 3 ' e 1 e2 4.3 C o r o l l a r y . I f _/\ i s a l a t t i c e of exceptional type, then j\ can be described as follows: i n case (a), A* = A L^(A + 0L-J2); . ep ep ep 0 • i n case (b)(7) A*p = A e p U (A0p+ 0^/2) and i n case (kHz*/) A*p = A>P> p = 1»2-// 4.4 Proposition. Suppose A i s a l a t t i c e i n v a r i a n t under K +. Then there e x i s t s a root system R1 such that W(R') = K and A i s eith e r of exceptional type w.r.t. R» or else QlRMC A C P ( R ' ) and A = Q( R ,)» In the l a t t e r case. A i s i n v a r i a n t under K and A* = P(R'), unless R« i s of type A. x ... * A. and A= A , when A* = P(R') U (P(R« )+(£" <* ) A ) . Proof. We f i r s t show that A ± t 0 f o r a l l o< € B., so that A , i s a l a t t i c e . I f dim V > 4, there w i l l e x i s t <x\,«r €B 2 2 d i f f e r e n t f rom o< • f o r w h i c h a., = 0. S i n c e A i s a l a t t i c e , we c a n f i n d an element x ^ A f o r w h i c h n e i t h e r o f ^x, and <^ x, <A .y i s z e r o . Prop.if. 1 t h e n shows t h a t j\. ji 0» When dim V = 3, t h i s argument works f o r a t l e a s t one o<_£B. I f ao<^ i s a n o n z e r o element o f A ^ and o*.. CB i s c o n n e c t e d t o , t h e n (1 + s .s. ) ( a o*. ) = n. .a ^ . i s a non z e r o element o f A . - . The o n l y r e m a i n i n g c a s e i s t h a t o f a r o o t o t ^ ^ B w h i c h i s n o t co n n e c t e d t o e i t h e r o f t h e o t h e r two r o o t s , &, € B, b u t f o r x 3 w h i c h we may assume t h a t n ^ = - 1 . L e t > f D e s u c h t h a t <x, <*k/ / 0. S i n c e z = '£if-n. .n.-) <(x, oOc* £A by p r o p . l f . 1 , i t f o l l o w s t h a t ( l + s ^ s^ X z ) = n i j ( ^ " " n j ; - j n - j ^ <*-j> However, s i n c e x - s ^ C x ) = < x , 1 ^ ; j , > ^ + <x, ^ k > ° ^ k ^ A , we deduce t h a t n ^ U - n ^ n ^ ^ x , e ? k > oi fc a l s o b e l o n g s t o A • T h e r e f o r e A ^ . ^ O. Suppose A ^ = Z m ^ o ^ f o r some > 0. S i n c e O-s . - X m.^o^) = (1 + s ^ s i ) ( m i c < i ) = '-o^ jBijL ^ ^ A ^ , A * i s i n v a r i a n t under K and t h e r e f o r e e q u a l t o Q(R') f o r some r o o t s y s t e m R' s u c h t h a t W(R») = K by 1 . 3 . 2 . A p p l y i n g prop.if.2, we see t h a t e i t h e r A CJ\ = P(R') o r A i s o f e x c e p t i o n a l t y p e w . r . t . R 1 o r R' i s o f t y p e A. X ... X A , and A = A „ . I n t h e l a s t c a s e , t h e v a l u e o f j\* l i e ' e a l s o f o l l o w s from p r o p . 4 . 2 . / / 5. The g r o u p s H 1 (K + ,V /_A) . Suppose A i s a l a t t i c e i n V = © Vg i n v a r i a n t under K + . Then A . c o n t a i n s t h e s u b l a t t i c e A ^ =. A K ® A ^ and t h e p r o j e c t i o n o f A on V £ is c o n t a i n e d i n A * . We may assume t h a t A K i s a s d e s c r i b e d i n p r o p . i f . b u t w i t h R 1 = R. I t w i l l be c o n v e n i e n t t o i n t r o d u c e t h e number (5) eij = mij/^-aijV 23 for tt., o(.£B. I t equals §,1,2 or 6 according to whether i s disconnected from C< or connected to i t by a s i n g l e , double or t r i p l e bond. Choose CX Q to be connected to at most one other root Oi^ and l e t G = {g^ = S ^ S Q j i / oJ be the corresponding set of generators of K + . Suppose t i s a function G — ^ V . I f t(g^) = y i + ^ k i ^ k ' w i t h y i € v K » d e f i n e u i j = < t < g i ) , ^ j > = 5.1 P r o p o s i t i o n . The function t induces a coboundary t:K +—>V/LA. i f f there e x i s t s a constant c e IR such that t C g ^ = c otQ + t ^ C X ^ mod_A for a l l i /0. Proof. I f a = y + ^ k ^ k ^ k ^ ' w i t n V t n e co° o uhdary corresponding to a mod A maps g^ to i - g ^ t = < Z " k V k 0 ) * Q + ( 1 ^ - B Q iZ^^o) <x± mod -A. Conversely, suppose t(S^) = c (x^ + t j ^ 0 ^ mod A. The system of equations ^ k W o = c Z k W i " n o i z k \ \ o " * ! ! can be written i n the form aH. = t, where a = (a^) , K i s the Cartan matrix of R and t i s the vector with 0-th coordinate equal to c and i - t h coordinate, f o r i / 0, equal to t - y / ^ o i 0 . Since H i s i n v e r t i b l e , the system has a unique s o l u t i o n a and "t coincides with the coboundary defined by a = ^ " ^ a ^ ^ ^ * / / I f e x i s t s , l e t c and d be the solutions of the system of equations n Q 1 c + 2d = u n 2c • n 1 0 d = u 1 Q ; 24 o t h e r w i s e , l e t c = 0. By s u b t r a c t i n g f rom t t h e f u n c t i o n gf —> + g±__•> c c ^ 0 + ^ i ^ ^ C M D we may assume t h a t u ^ = 0 f o r a l l i and a l s o t h a t u ^ Q = 0. 5.2 P r o p o s i t i o n . The f u n c t i o n t i n d u c e s a c o c y c l e ~t:K*—5> V / A i f f the f o l l o w i n g c o n d i t i o n s a r e s a t i s f i e d f o r a l l i , j " 4 0: (6) 2 t ( g ± ) = u i Q < * 0 mod A i f i / 1 (7) * l 0 t ( 6 i > 6 -A-(8) ^ ( t C g ^ - t C g ^ = • i J f - < 2 » a i * a 3 1 « 1 j ^ i + ( 2 t tir n ij uji^j] m o d P r o o f . I f h£ K + i s o f o r d e r m, we s h a l l d e n o t e 1+h+.. ,+hm" 1 by 31(h). Our f i r s t a im i s t o e s t a b l i s h t h e f o r m u l a (9) N ( s i s . ) t ( g i ) = m ^ t C ^ ) • e ^ U ^ o ^ - 2 u . . c V j ) f o r i 4 0. I n d e e d , s i n c e s ^ ^ S ^ ) = t ;^Si^~uij ^j+njiuij ° * i t y (2), t h e l e f t s i d e o f (9) e q u a l s m ^ i C t C g ^ - v ) , f o r some v i n Ej^t S i n c e t h e l e f t s i d e o f (9) i s a l s o i n v a r i a n t u n d e r s ^ s ^ » t h e v e c t o r v must i n f a c t be t h e o r t h o g o n a l p r o j e c t i o n o f t ( g i ) on t h i s p l a n e . C o n s e q u e n t l y , i f v = v . + v . <X . , we have the e q u a t i o n s 0 = u ± i = < t ( g i ) , \ > = < v , ^ ± > = 2v ± + n ^ uij = < t C g i ) , ^ i > = < v , ^ > = n ^ + 2v. , whose s o l u t i o n s a r e v i = - ^ / ^ J V V J = 2vu-*ijnji} ; t h i s e s t a b l i s h e s 19). A c c o r d i n g t o the p r e s e n t a t i o n CO o f K + , t h e f u n c t i o n t w i l l i n d u c e a c o c y c l e " t : K * — * V /A i f f t i S j L ) = N ( g i ) t ' ( g i > = N ( s i s 0 ) t ( g i ) ^A. and tCCgj.g]1) i 3 ) = H C g ^ K t C ^ ) - s±s1}Ug.))e A . Since g.g" 1 = s. s. and H(s.s.) = B{s.s.), the l a t t e r r e l a t i o n x. 3 x J X J J X can also be expressed as M (s j.s^)t(g i) = HCs^s^tCg^) mod A . Using 19), one sees that a r e l a t i o n of the f i r s t type amounts to (6) or (7), while that of the second type amounts to (8).// We s h a l l postpone u n t i l the end of t h i s s e c t i o n the discussion of the cases when A„ i s of exceptional type or when R i s of XV type A * . . . * A, and Au = A . This enables us to assume i l xv e that A * = P(R), so that oL±iA) C. TL for a l l oC± < B. 5.3 Proposi t i o n . For a l l i , j / 1 , we have (aj 2 u i 5 £ P. (b) 2u. 1 u a 0 1 u i 0 mod / ; m 1 0 u 1 1 6 1'. (c) 2 i * i 0 = 2u.-0 mod Z i f m±^ = 2 or if; 5u±Q = 3^0 mod 21 i f = 3 or 6. Proof. Evaluating *L . at (6), we obtain i a ) ; to obtain ( b ) , J evaluate and <x.^ at (6) and (7) r e s p e c t i v e l y . v _ I f = 2, evaluating <X-Q a t (8) shows that 2 ( u i o " u j O ^ * S i m i l a r l y , i f = 3, evaluating UQ at 18) shows that 3(U;J.0-U.J0)£ ^ . I f m.^ = if or 6, then i s even and therefore the r i g h t side of (8) i s i n A by (a) . Using (6), we conclude that. -Jm^^ ( H^Q-U j Q) oCQ £ A, which can only happen i f i m ^ t u ^ - u . ^ ) ^ J T . / / I f R i s of type B, XA, x ... * A , , we define A to be the 3 1 1 e su b l a t t i c e of P(R) which contains Q(R) and a l l elements ( o * Q + o<2+ ^^)/Zy where oiQ and o<2 are the f i r s t and t h i r d roots i n the component of type B^ and i s i n a component of type A j . 26> ( I f there are no factors of type A., l e t Aa = Q(R).) I e 5.§ Proposition. We have ^ - ^ i ^ i o ^ O € % for a l l i ^ 1, unless (a) n l Q = -2 and n ^ = 0, when * ± 0 g Z . (b) OLQ belongs to a component of type A^.D^.B, or B^(in the l a s t two cases., c* 0 i s the f i r s t and the t h i r d root) and (<*0+ *L%)/2 6 A , when 6u i Q£ Z. (c) R i e of type X-Aj * ... * A 1 and - A £ x A e , when 6u 2 Q6 Z . Proof. Eqn.(8) with j = 1 gives OO) m i l ( t ( g i ) - t ( g 1 ) ) = e . 1 [ - ( 2 u l i + n l i u i l ) ^ i + ( 2 u 1 1 + n : L 1 u l i W 1 ) mod A . . Evaluating ^ Q at t h i s equation, we deduce that (11) m i 1 u i 0 = e i l n 1 0 ( 2 u 1 1 + n i 1 u 1 i ) mod Z . Suppose = 0. I f » 1 Q = -2, (11) shows that 2u i Q = -211^ mod ff; since n Q 1 = -1, 5.3(b) implies that u i Q £ Z. Otherwise, multiplying <1l) by 2 and using 5.3(b), we conclude that ( ^ " n o i a i 0 ^ u i O £ Z . I f n ^ 0 and n 1 Q = -2, we have = 3, so that (11) becomes the equation 3u^0 = ~ * * u j j + 2 u i i m o d & • M u l t i p l y i n g t h i s by 2 and using 5.3(b), we again conclude that (if-nQ 1n^ 0)u^ 0£ Z» Suppose n ^ / 0 and n 1 Q = -1. M u l t i p l y i n g (11) by m 1 Q and using 5.3(b), the conclusion can be expressed as 2(4-n 0^n^ 0)u i Q € Z . I f n Q 1 = -2, we see from 5.3(b) that both kVL^^ and Z f U ^ are i n Z . M u l t i p l y i n g (10) by Zf and using (7), we deduce that 12t(g i ) 6 J A ; i t therefore follows from (6) that 6uiQ€ Z and hence 2u j L 0 = ( H 0 1 n 1 0 h i 0 ^ . I f n Q 1 = -1 and m^ s 3, 5.3(b) implies that 12uil and 3^^ are i n Z • M u l t i p l y i n g <10) by 6 and using (7) and 5.3(b), we deduce that l8t ( g i ) = -3^0 ^ m o d i n view of (6) and the fact that 6u i Q€ 2?, t h i s means that 3 t t i 0 ( . V" o t i ) T h u s e i t h e r 3 u i 0 = ( Z f ~ n O l n i O ) u i O € ^ o r 27 ( c<0+ ot~)/2. € A . However, one can easily see that ( ^)/2 can be a weight only i n the cases specified i n (b). Finally, suppose n(^ 1 = -1 and = 4. I f n^ = -1, multiplying (10) by 3 and using 5.3(e)* (6) iUnd (7), we see that 3u i 0«' i^A, so that 3 u i Q = ^ ~ a o i n i 0 ^ u i O € ^ * T n e r e m a i n i n S case i s when c<0 and o<^ are the f i r s t and third roots i n a component of type B^. If R i t s e l f i s of type B^ and (ot Q+ ^±)/Z A. R must be the l a t t i c e Q(R) = A . Otherwise, suppose . i s not i n this e j component; eqns.(6) and (8) show that (u. ~-u ._) „ = - u c ^ - . + u. , . mod -A- . x 10 oO 0 J i i w i j Q * Multiplying this equation by 3 and using 5.3(a) and the fact that 3UJQ£ Z„ we deduce that 3 u i o ( > ' o = u j i °*i + u i j m o d Evaluating oL^ at this equation shows that u ^ = 3uJ_Q» I i 3 u i 0 ^ 1 and (o* Q+ o*±)/2 ^ A, then we must have (c* Q+ ^ i+<KJ)/2^ 'A However, such an element can only be a weight for every j only i f (joi .] i s of type A.; this produces the exception noted i n (c), since A i s a maximal sublattice of P(R) w.r.t. the property (A e) # = Q(R).// 5.5 Proposition. 6.H1(K+,Y/A) = 0, unless R i s of type A, X ... * A,, B_ *A, X ... X A, or B, XA..X ... XA, and, i n the 1 1 d 1 1 "~~ 3 • I . last case = A g . In the last two cases. 6H (K ,V/A) i s of order ^ 2 . Proof. Suppose B contains a component of type 6^. I f we choose vlQ to be the f i r s t root of such a component, 5.4 says that U i 0 6 ^ f o r 3 1 1 1 ^ U so that 2t(g ±)£A by (6). In view of (7), we have 6t » 0. Secondly, suppose B contains a component of type d O other than A and Bg and choose o£ Q to be the f i r s t root of such a component; then mQ^ = 3» We s © e from 5 . 4 that 3u^ Q £ Z for i / 1,, and thus by ( 6 ) , 6 t ( g ^ £ A, apart from c e r t a i n exceptional cases,, when 6t(g±) = c* 0 /2 mod A . I f ( oiQ+ *±)/2. «A, then S U g ^ ) = <*.j/2 mod A; since 3 t(g t)*A by ( 6 ) , prop . 5 . 1 shows that 6t i s a coboundary. The only remaining exception occurs when R i s of type B^* A^ x ... X A^ and -A^ = A e ? F i n a l l y , suppose B contains a component of type Bg. Choosing Q to be the second root of such a component, so that n^ Q = - 2 , we see from 5.1+ that; Q 6 7Z for a l l i / 1, and thus 2t(g i)<£ A. Eqn . ( 8 ) with j = 1 then says that (12) 2 t ( g i ) = u l t^ t - u i lcyL1 mod A. Since J+t^g^eA by (7) and 2 u i l«Z by 5.3(b), we must also have au.. 6 I f there e x i s t i , j ^ 1 such that n. . = - 1 , apply-in g d&tj to (12) shows that \x^€"E then 2 t(gj) = - u i l c X 1 and 2t i s a coboundary. This f a i l s only i f R i s of type B ^ X A ^ ... ^ A ^ when we may assume that 2t-'(g^) = oC^/2 mod A f o r any i / 1. The only Remaining case i s when R i s of type A^y. ...X-A^// In each of the excluded cases, one can f i n d examples of l a t t i c e s A for which 6H 1 (K+,V/_A-) / 0 . For instance, suppose R i s of type Aj X-A^ with 5 f a c t o r s and that .A contains (0*2+ * y / 2 , ( ^ + ^ 5 ) / 2 and ( c*0+ <Xg+ ^ ) / 2 . Define t(gg) = t(g 3> = (<*0+ * + ^ 5 ) A and t(g^) = t ( g 5 ) = (oCg+ * ) A ; then t i s a cocycle but 6t i s not a coboundary. 5 . 6 Proposition, i+.tt' (K+,Y/A ) = 0 , unless R i s of type A, x A , A,XG 2, A 3 , A , x A 3 , A 2 X A 2 , A 2 x G 2 , A^, D^, A ^ x ^ , A,XD^ , A_. D^ or A. x D,.. 29 Proof. Suppose R contains a component of type Bffl or C m f o r some m ^ 2 . Let ©< Q be the l a s t root i n such a component. By 5«4* we have Zn±Q6 ^ f o r a l l i / 1, so that 2ft ( g ^ ) 4 A . From (7), we also have A , and thus 4t = 0. Secondly, suppose R contains a component TT of type I f c^0 i s the f i r s t root i n TI , 5.4 shows that u i Q € 2? f o r a l l i / 1, so that 2t(g^)€-.A . Eqns.(6) and (8) together imply that for a l l k / 1, (13) 2t( g ]) . -u^ + u l k * k mod A, Suppose there e x i s t two disconnected roots ., ^ . ^BNTT. L e t t i n g v k = j i n 03) and applying ck^, we conclude that 2M^€ "Z • Since 2u i l £ "E by 5.3(b), we see from (13) with k = i that 4~t = 0. I f B \ H" = i s of type G,, t h i s method of argument shows that 2u^ = - 3 u ^ mod H and 2u^ = - u ^ mod 2? \ together, these equations again imply that 2u^ £ %. The only other p o s s i b i l i t i e s are A ^ G 2 and k^yC Otherwise, i f cx^ e x i s t s , we have = 3. Exclude f o r the moment the cases F^ and Dg i n addit i o n to the l i s t e d exceptions. Let B« =(^±6-B 1 i i 0,l} ,and choose some o/^£B' which i s connected to at most one other o^^B'. I f b ^ ^ B ' i s d i f f e r e n t from «*• 5.3(c) says that 2u i Q = 2u k Q mod . On the other hand, our assumption implies that there e x i s t s ^ t B 1 disconnec-ted' from <y-.., so that 2u^ = Za^ = Zn±Q mod Z% Thus the numbers 2u i Q have, mod 2? , a common value c for i / 1. I f bl^ i s disconnected from f e q n s > ( 6 ) and (8) show that (14) 2t( g l) = u w 0 « 0 - u ^ o ^ + u ^ c ^ mod «A. 30 Again, our assumption implies the existence of two elements o^, i n B 1 disconnected from ^ and ffom each other. Applying ©C i to (14) with m = j shows that 2.u^€ u . Then, on m u l t i p l y i n g (14) )>y 2, with m = i , we deduce that 4 t(g^) = e c > t o ~ 2 u i 1 °S mo<*A ? Since 4t(g ; i L) = c ^ Q mod A f o r i / 1, 5i\ sKows^that 4*t i s a coboundary. The cases and Dg w i l l be handled during the next argument.// Since there i s no overlap i n the exceptional cases of the preceding two propositions, we conclude that either 4 or 6 always a n n i h i l a t e s H 1 ( K + , V / A ) . The group 4 H 1 ( K + , V / A ) i s therefore an n i h i l a t e d by 3 i n exceptional cases of prop.5.6.The exact sequence 0 - ^ V A - ^ V / A - ^ V / A - > 0 , where A * = V K © F ( R ) , gives r i s e to the exact sequence Hom(K+, A * / A ) ^ > H 1 (K +,V / A >->H1 (K +,V K/P(R)) . 1.2 shows that unless R i s of type A^ or D^, the f i r s t group i n t h i s sequence has no 3-torsion. Thus the 3-torsion component of H 1(K*,V/ A ) maps i n j e c t i v e l y i n t o that of H 1(K +,V K/P(R)). When R i s of type *^ or Dg, - 1 y <£ K + and 1 + ^ therefore H (K ,V K/P(R)) i s a n n i h i l a t e d by 2. Otherwise, one fi n d s by e x p l i c i t computation that the 3-torsion component of E 1(K +,V K/P(R)) i s isomorphic to Z / 3 Z i n a l l cases except that of AgXAg, when i t i s isomorphic to (2V3jP)^» When R i s of type &^ or D^, the 3-torsion component of H 1(K +,V K/P(R)) vanishes, but the f i r s t group i n the sequence i s isomorphic to the group of elements i n (V s Q P ( R ) ) / A an n i h i l a t e d by 3. We thus have i n a l l cases a bound for the dimension of 4H 1(K +,V/A) over 2?/3 Z. 31 As K + i s normal i n K and A i s thy assumption) i n v a r i a n t under K., we have the exact sequence 0 - > H 1 C K A + , ( V M ) K + ) i ^ H L ( K , V / A ) £ 2 ^ H 1 ( K + , Y / A > Since K A + acts t r i v i a l l y on ( V / A ) K = ( V ? © P ( R ) ) / A * the f i r s t group* can be i d e n t i f i e d with the group of elements i n ( V ^ + P ( R ) ) / A a n n i h i l a t e d by 2. On the other hand, since 2 H 1 ( K , V / A ) = 0 by 1.3.3, the image of res i s contained i n the subgroup H 1 ( K + , V / - A ) 2 of elements a n n i h i l a t e d by 2. 5.7 Proposition. The image of res i s of index 1 or 2 i n H 1 ( K + , V / _ A ) 2 . In the l a t t e r case, the components of R can only be of type A 1 or Ay Furthermore, i f R Q i s a component of type Ay then Q ( R Q ) C A n P C R Q ) C P d y . Proof. Suppose ~t i s a cocycle J t + — > V / A such that 2t i s a coboundary, i . e . -(15) 2tig ±) = coC Q + mod . A , for some c,d^ € /R. We s h a l l f i r s t show that one can, assume c = 0. I f R i s of type Aj * ... XA^, we can subtract from t the coboundary inducing function g^—^i-coC^, which does not a f f e c t the assumption that % . Otherwise, we can choose o i Q such that Cx^ e x i s t s and n Q 1 = -1. Evaluating <*Q and oi 1 at (15) with i = 1, we conclude that Zc+n^d^ and -c+2d,j both belong to Z . I f n l Q = -2 or -3, i t follows that c c S.» I f n l 0 = -1, then 3c € £ and d^ = 2c mod ~E. Therefore, adding to t the coboundary inducing function g^—=> c <xQ ( i / 1), g 1—=>cc^ Q + 200^ does not a f f e c t the assumption that VL±± and U10 a r e i n ~^ * t u t e n a b l e s u s to assume that c = 0. The argument also permits us to assume that d, =0, unless n,„ = 32 when d 1 = 0 o r i . E v a l u a t i n g c<^ a t (15)> we c o n c l u d e t h a t d^^ can be assumed t o be e i t h e r 0 o r i. Eqn.(6) a l s o shows t h a t f o r i / 1, u i 0 = d i m o d ^ a n d <16) d i = i i m p l i e s < °<Q+ o*±)/Z € A . I f n 1 Q = -3, t h e n u i Q e J by 5.4 and t h u s d ± = 0 . I f n^Q = -1 and we a r e n o t i n c a s e (bj> o f 5.4, i t a g a i n f o l l o w s t h a t d i = 0 . I f O ( Q i s t h e l a s t r o o t i n a component o f t y p e B 2, t h e n d i = 0 f o r i / 1 s i n c e (tf Q+ ^ . ^ / S c a n n o t be a w e i g h t . ITg s u c h t h a t < x , ~ Q , 5.8 Lemma. Suppose t h e r e e x i s t s an x £ V K s u c h t h a t <x, <^ ^ ^ * 2 <^x, 6 J , < x , = d± mod Z l o r i ^ 1 and 2x £ A . Then t h e f u n c t i o n T ( s Q ) = x , T ( s j L ) = tig/) - x - £ ^ 6 ^ ' + a c * 0 , where a = d^ - <x , ^ ^ , i n d u c e s a c o c y c l e T : K — » > V / A s u c h t h a t the r e s t r i c t i o n o f T t o K + i s cohomologous t o * t . P r o o f . Note t h a t p Q 0 s ^ x , o^ Q> € Z and p ^ = \x^± - ^ x , * ? ^ + d^ + a ^ Q i 6 % f o r a l I i » t h e r e f o r e , we have o n l y t o v e r i f y t h e c o n d i t i o n s <3)-(5) o f I.3.3. I t i s c l e a r t h a t 2 T ( s 0 > s 2x €J\ and 2 T ( s ± ) eA f o r a l l i . I f i 4 1, P Q i = d± mod Z and p i Q = u i Q mod Z , so t h a t p Q i c*^ = P i 0 m o d " ^ " * I f ^ and a r e d i s c o n n e c t e d r o o t s i n B d i f f e r e n t f r om oC Q , we have p... = u ± f d y P ^ = and - d . ^ = u ^ . . - u ^ o ^ mod A by <8), so t h a t p ^ ot = P j 1 ° t i mod A I f m 0 1 = 3, t h e n 3t (g j )«A by (7) and 2 t ( g t ) c A by (15), so t h a t t (g 1 )<sA. T h e r e f o r e p Q l = a mod Z , P 1 0 £ J and T ( S l ) . T ( S o ) = p 0 1 o t o . p 1 0 (* t mod A . I f o t ± and ^ a r e d i f f e r e n t f rom and a r e s u c h t h a t m. . = 3, t h e n p. . = \ $ " d j + * d i m o d ^ » p j i = u j i " d i + m o d & • A p p l y i n g 33 . to (15)» v/e have 2u. . = -d. and s u b s t i t u t i n g t h i s i n (8), we obtain t(g±) - t(g..) = ( d ; f u i j * ( * i - " ' * d i ~ u ; J i ^ J m o d - ^ - » so that TCs^^} - T(Sj) = - P ^ O ^ mod A . Therefore T induces a cocycles • 7c :.&—-> V/A ; since T(g A) = TCs.^) + S ^ C S Q ) = "tCg.^ + ao< 0 - <-Jdj. + ^ x , ^ . ^ )<* i mod A , the r e s t r i c t i o n of T to K + i s cohomologous t o ~ t . / / Returning to the proof of the p r o p o s i t i o n , suppose d^ = 0 f o r i 4 1; we can apply 5»8 with x = 0 to reach the conclusion. I f oLQ and are the f i r s t and t h i r d roots i n a component of type or B^ and i f d^ = -Jr, we can l e t x s -fe^g, since by (16). The same kind of argument works for a component iQ of type A ^ or D^ i f A H P C R Q ) = P ( R Q ) . In case RQ i s type and A O P(RQ) =- q(RQ) U (Q(RQ) + for some j , we can choose o(Q = ot^ . and again have d^ = 0. There remains the case when a l l components of R are of type A 1 or A^. Define a homomorphism ^r: H 1{K +,V/A) 2->i7/2 by C ( t ) = X . . / , d ' mod Z . I f c r(t) = 0, then d , = 0 i n case there i s a component of type A^» so that the previous argument applies, or the number of d i = i i s even i n case R i s of type A j * . . . x* Aj • However, v/e can then l e t x = 2l <^ ^ i n 5»8 since 2x = Z d i^^^-A i n view of ( l 6 ) . / / One can give examples i n which the image of res i s a c t u a l l y of index 2 i n H 1 (K+,V/A)->» Suppose f i r s t that R i s of type A, and A i s the l a t t i c e A g J then the function t(gj') = 0, t(gg) = 0*.|/2 + c 2^/4 provides a counterexample. Secondly, suppose R i s of type k^ ... XA^, with k f a c t o r s , and A contains (^Q-*- °^ 34 and (<* 0+ &^)/2; t h e n we c a n d e f i n e t ( £ g 2 ) = ( &Q+ + c>c2f)A and t ( g ^ = t ( g ^ ) = 0. I f we e x c l u d e a l l the e x c e p t i o n a l c a s e s o f the p r e c e d i n g 3 p r o p o s i t i o n s , we see t h a t H 1 ( K + , V / A ) i s a n n i h i l a t e d by 2 1 1 + and t h a t t h e r e s t r i c t i o n honiomorphism H ( K , V / A ) — > H (K , V / A ) i s s u r j e c t i v e . G i v e n a c o c y c l e t : K+—"> V / A , one can t h e r e f o r e f i n d a c o c y c l e T:K—=>V//\ s u c h t h a t T J g + = t . A space group w i t h p o i n t group K + c o r r e s p o n d i n g to t c a n t h e n be v i e w e d a s t h e s u b g r o u p o f d i r e c t i s o m e t r i e s o f a space group w i t h p o i n t group K c o r r e s p o n d i n g t o T . F o r l a t t i c e s o f e x c e p t i o n a l t y p e , e x p l i c i t c o m p u t a t i o n shows t h a t i n case ( b ) , H 1 ( K + , V / A ) = I ~Z/Z ^\tw©(A). = 0 a n d ( Z / 2 Z ) d i m / ' w K " 1 i f (9(A) = w h i l e i n c a s e ( c ) , H 1 ( K + . , V / A ) = iJ/Z I ) 6 1 * 1 We s h a l l n o t d e a l w i t h t h e c a s e when R i s o f t y p e A j X . . . * A j s i n c e t h e s i t u a t i o n i s n o t c l e a r even i n t h e o r d i n a r y c a s e . 6. A p p l i c a t i o n t o c l a s s i c a l c r y s t a l l o g r a p h y . I n t h e two and t h r e e d i m e n s i o n a l c a s e s , t h e above r e s u l t s e n a b l e one t o g i v e a ' t h e o r e t i c a l ' d e r i v a t i o n o f some space g r o u p s w e l l - k n o w n i n c l a s s i c a l c r y s t a l l o g r a p h y . I n two d i m e n s i o n s , t h e r e i s a l w a y s o n l y one e q u i v a l e n c e c l a s s o f l a t t i c e s and the group H (K , V / A ) i s a l w a y s z e r o . The t h r e e d i m e n s i o n a l s i t u a t i o n i s p r e s e n t e d i n T a b l e I , u s i n g t h e S c h o e n f l i e s n o t a t i o n f o r ease o f c o m p a r i s o n w i t h j . The e n t r i e s f o r g r o u p s o f t h e form +K*" were deduced f rom 1.2.5. 35 Table I K + A © < A ) C l a s s i c a l H ^ K ^ Y / A ) ^ o r b i t s Q ( R ) 0 1 2 V 2 2 2 P ( R ) ** ^/2 Z 2 Q ( R ) * * ' 3 P ( R ) o r Z/zZ 2 'Q(R). ° P 1/zZ 2 Q ( R ) D3c7 Z / 3 Z 2 Q ( A 1 ) + P ( A 2 ) ha 2 7 3 Z 2 A , n 0 1 el i ( A 1 X B 2 ) + \ ^ Z/K7ZG1/ZZ 6 { (^ 1 +^ 2)./ 2] 2?tZZ 2 ( A 1 X G 2 ) + < Q ( R ) D6 Z/Zl® 273Z ( A 1 X A 1 >< A 1 ) + Q ( R ) Z/Z1@T/Z1® T/Z2Z. 4 7Z/ZZ 2 Z/Z7Z 2 0 1 Q ( R ) 0 2 7 3 ^ 2 Q ( R ) P ( R ) / Q ( R ) 0 1 <Q( R ) 0 3 + q(R) P ( R ) / Q ( R ) ^ / 2 Z 2 Q ( R ) 0 ° 6 Z/Zl®!/?,? 4 H Q ( R ) 0 C20( Z/z-z 2 Q ( R ) { o , ( c / 1 + 6i2)/23 0 1 A1 any °1 0 1 36 T a b l e I ( c e n t ) as f o r At T -r>, 3 3 Thp 272 Z 2 t A 2 a s f o r A * ^ . j Q ^ C3 i * 0 1 +B+ a s f o r B+ l/ZlQI/zV f k +G+ a s f o r G+ ? / 2 Z 2 +A. a s f o r A t C . 0 1 — I i i I n t h e above t a b l e , we have abandoned o u r c o n v e n t i o n r e g a r d i n g oiQ and 1 and s i m p l y numbered the e l e m e n t s o f B c o n s e c u t i v e l y s t a r t i n g w i t h oi . n o t a t i o n s u c h as d e n o t e s t h e l a t t i c e Q(R) U ( Q ( R ) + t O ) . 37 BIBLIOGRAPHY 1 . N.Bourbaki, Groupes et algebres de L i e , Chap.IV-VI, Hermann, P a r i s , 1 9 6 8 . 2 . H.Brown,J.Neubuser and H.Zassenhaus, On i n t e g r a l groups,I, Numer.Math. 19.(1972), pp.386-399. 3. i b i d . . On i n t e g r a l groups I I , Numer.Math. 20(1972), pp.22-31 4. i b i d . . On i n t e g r a l groups I I I , Math. Of Computation, 22(1973), pp. 167-182. 5. R.Bulow,J.Neubuser, On some a p p l i c a t i o n s of group t h e o r e t i c a l programmes to the de r i v a t i o n of the c r y s t a l classes of R^, Computational problems i n abstract algebra (Proc.Conf.,Oxford, 1967), Pergarmon Press, Oxford, 1970, pp.131-135. 6. J.Burkhardt, Die Bewegungsgruppen der K r i s t a l l o g r a p h i e , Birkhauser, Basel, 1 9 6 6 . 7. R.Carter, Simple groups of L i e type, Wiley, 1972. 8 . H.S.M.Coxeter and W.Moser, Generators and r e l a t i o n s for d i s c r e t e groups, Springer, 1 9 7 2 . 9. E.Dade, The maximal f i n i t e groups of 4*4 matrices, I l l i n o i s ff.Matk. 2(1965), pp .99-122. 1 0 . C.Hermann, K r i s t a l l o g r a p h i e i n Raumen b e l i e b i g e r Dimensionszahl, Acta Cryst. 2(1949), PP.139-145. 1 1 . A.Hurley, F i n i t e r o t a t i o n groups and crystal'.classes i n 4 dimensions, Proc.Camb.Phil.Soc. 4JZ.095D, pp . 6 5 0 - 6 6 l . 1 2 . G.Maxwell, The crystallography of Coxeter groups, J.Algebra, to appear 1975. 13. S.Ryskov, On maximal f i n i t e groups of integer n * n matrices, Dokl.Akad.Nauk SSSR, 204C1972). no.3,; t r a n s l a t e d i n Soviet Math.Dokl. !2(1972j>, pp.720-724. H.Weyl, Symmetry,Princeton, 1952. H.Zassenhaus, Uber einen Mgorithmus zur Bestimmung der Raumgruppen, Comment. Math. Hei v . 2J.( 1948), pp. 117-1J+1.
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- The crystallography of the rotation subgroups of Coxeter...
Open Collections
UBC Theses and Dissertations
Featured Collection
UBC Theses and Dissertations
The crystallography of the rotation subgroups of Coxeter groups Broderick, Norma 1975
pdf
Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
Page Metadata
Item Metadata
Title | The crystallography of the rotation subgroups of Coxeter groups |
Creator |
Broderick, Norma |
Publisher | University of British Columbia |
Date Issued | 1975 |
Description | The theory of space groups has its origins in crystallography and solid state physics. In this thesis, we study those space groups in n-dimensional Euclidean space whose point groups are the rotation subgroups of crystallographic Coxeter groups. |
Genre |
Thesis/Dissertation |
Type |
Text |
Language | eng |
Date Available | 2010-01-27 |
Provider | Vancouver : University of British Columbia Library |
Rights | For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use. |
DOI | 10.14288/1.0080110 |
URI | http://hdl.handle.net/2429/19184 |
Degree |
Master of Science - MSc |
Program |
Mathematics |
Affiliation |
Science, Faculty of Mathematics, Department of |
Degree Grantor | University of British Columbia |
Campus |
UBCV |
Scholarly Level | Graduate |
AggregatedSourceRepository | DSpace |
Download
- Media
- 831-UBC_1975_A6_7 B76_3.pdf [ 3.09MB ]
- Metadata
- JSON: 831-1.0080110.json
- JSON-LD: 831-1.0080110-ld.json
- RDF/XML (Pretty): 831-1.0080110-rdf.xml
- RDF/JSON: 831-1.0080110-rdf.json
- Turtle: 831-1.0080110-turtle.txt
- N-Triples: 831-1.0080110-rdf-ntriples.txt
- Original Record: 831-1.0080110-source.json
- Full Text
- 831-1.0080110-fulltext.txt
- Citation
- 831-1.0080110.ris
Full Text
Cite
Citation Scheme:
Usage Statistics
Share
Embed
Customize your widget with the following options, then copy and paste the code below into the HTML
of your page to embed this item in your website.
<div id="ubcOpenCollectionsWidgetDisplay">
<script id="ubcOpenCollectionsWidget"
src="{[{embed.src}]}"
data-item="{[{embed.item}]}"
data-collection="{[{embed.collection}]}"
data-metadata="{[{embed.showMetadata}]}"
data-width="{[{embed.width}]}"
data-media="{[{embed.selectedMedia}]}"
async >
</script>
</div>
Our image viewer uses the IIIF 2.0 standard.
To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0080110/manifest