This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
#
x
Let
A
A x E A2". zP s o t h a t e i t h e r "1 = xp
x1 E A2np
i s an element
f o r every
0
be a Hopf algebra over
x E A&
the dual of
(A*
with
A)
0.
m implication of
x = xoy xl,.
.. ,xm
so that
15 m
x E A2ny
5m
i s a sequence
i s a 1-implication of
xi+l
x
i'
Browder's f i r s t implication theorem (6.1 i n [ B r ~ w d e r ] ~ ) :
3.3.12.
Let
X,p
be
~n
homology Bockstein
Ery
H-space,
8.8.
I f
Br t h e Bockstein x = B ry
x E PEgn,
s.s,
Er
i t s dual
then x has m
impzication. Outline of proof:
... E Er
xo = x, xl, x2,...,xm, xi
sequence
xi E PE 2np
r
i-1
the
i
,
ro = r,
by a 1-implication of
xi's
to
with
xi
a 1-implication of
i s constructed inductively as follows.
sequence
r
One has t o construct a sequence
/-'= Er
x
0
i-1
5 ri and
- ri-l x
i
i-1'
i s represented i n
One can p u l l b a c k
t o obtain the desired implication sequence.
t h i s inductive argument it s u f f i c e s t o construct
x
1
The
One obtains a
5 1, xi
r = B iyi.
x.
out of
x = x
0'
By
97
Browder's Bockstein s p e c t r a l sequence
If
2=
XP = + + o
<X,]L> #
and suppose
0
x1 =
put
0
prove now t h a t
B
3.2.8 and 3.2.10
{?I
<{a , Br+l{XP-ly}>
XP =
r1 = ro,
-x E .:E
Br(xP-'y)
# 0, hence, =
r+l
2,
{p-' - Bry) #
?-$r
computing the
%rFy $-'p # 0 , hence one can conclude coproduct of 2 - l ~
consequently
x1 =
-
x,
in
z,x># 0 B
r
.
implies t h a t
x1 = B
r+l yl,
5
+o
so suppose
= xp = 0.
One can
-x # i m 8 , #
0 and
=
r+l {P-ly}
~,(Xp-'y).
0.
{xP-ly} # 0 that
and by
and
{XP-ly)
and
<{;;PI, xi> # o f o r
are primitive.
can be represented by a 1-implieation
y1 = {XP-ly>.
This method y i e l d s f u r t h e r r e s u l t s , a sample of which we present here
( f o r a more complete l i s t and references, see [ B r ~ w d e r ] ~ ) .
3.3.13.
Browder's implication theorems:
Bockstein s. s and
3.3.13.1.
If
X
x E P ( c ) Brx # 0 then x has an
P
m
,Z)
m
implication.
Hence
0.
If X is HC and p = 2 and if x E P ( E F ) then x has an
implication.
Hence, if H*(X,Z2)
particular PH*(X,Z2)
3.3.13.3.
be t h e H*(
an H-space.
if H*(X,Z ) is f i n i t e then BrP(EF) = 3.3.13.2.
Let {En,Bn)
If
X
Hence, if H*(X,Z
cH
odd
(X,Z2)
is f i n i t e then PE? and
H*(X,Z)
)
is f i n i t e
H*(X,Z)
in
has no 2-torsions.
is HC and HA then x E P E T has an P
= 0,
has no p-torsion.
m
implication.
98
The cohomolojg of H-spaces
3.4.
High order operations. The systematic study of high order operations s t a r t e d e s s e n t i a l l y i n
I n general, l e t
[AdamsIl.
be abelian groups.
G,G'
cohomoZogy operation i s a t r i p l e a = (x,E,y)
u i s i n t h e domain of The value
Given a space
y E H*(E,G').
x E H*(E,G),
a(u)
a
i f there exists
K
where
A (high order) E
i s a space,
and a c l a s s
f: K
-+
E
-+
E,
G-G'
with
u E H*(K,G), H*(f,G)x = u.
i s then defined as t h e set
a(u)
= {H*(f,G')y E H * ( K , G ' ) I f :
K
H*(f,G)x = u).
-
One can e a s i l y replace x by a f i n i t e sequence x = xly...,x m' n. nJ xi E H '(E,G), y = yl,.. .,yn, y E H (E,G') and analogously define t h e
-
J
domain of
a
and i t s values.
arbitrarily.
Suppose
operation as
(x,E,y)
I n practice
A EJ 2
f j : EJ-1
-+
KJ-1'
K
j
are not chosen
-
G = G' = Z One speaks of a Z Z k-order P P P where E i s a generalized k-stage Postnikov system:
1 mj
x,E,y
= K(G,)
P
is the f i b r a t i o n induced by where
G,
i s a graded
Z P
vector space.
99
High order operations
x
i s then chosen t o be t h e image of t h e vector of fundamental classes i n
= H*(K
H*(E1,Zp) all i n
i m H*(r
Z ) 0' P
i s assumed t o be a vector of c l a s s e s not
and y
Z 1. k' P
If all spaces and maps are
-
loop spaces and maps and y
i s an
t h e operation i s then c a l l e d s t a b l e .
-suspension,
I n t h i s s e c t i o n we i l l u s t r a t e a technique used i n t h e study of t h e
.
cohomology of H-spaces involving a secondary operation (i. e : an operation of order two).
We s h a l l give only the most elementary version.
A t the
end of t h i s s e c t i o n w e Bhall l i s t some of t h e stronger r e s u l t s obtained using t h i s method, as well as r e s u l t s concerning t h e cohomology of H-spaces obtained by using operations i n generalized cohomology theories. We s h a l l study only one secondary operation. Let
K(Z
induced by
P'
5 : K(Zp,2n)
x = H*(rl,Zp)i2n.
Proof:
hp-1)
-
Note t h a t
*
r
El
1 K(Zp,2n)
K(Zp,2np) H * ( 5 , Z p ) ~ 2 n p El
i s an
m
=
be t h e f i b r a t i o n Put
loop space.
Consider t h e s t a b l e diagram of induced f i b r a t i o n s :
The cohomology o f H-spaces
100
For m = 2n
one gets a diagram of
loop spaces and maps
lr
K ( Z p y 2n)
By 3.3.9
multiplication
QE
K ( Z p y 2n)
K(Zp,2n)
w
x
K(Zpy2np) ( t h e l a t t e r with a twisted
u,).
I n t h e notations of 3.3.9 one has an H-map g: E1,Add
K ( Z ,%) P t h e desired c l a s s . --*
Note t h a t
y
x
K(Zp,2np),
uV
and
y = H*(g,Zp)(l B
I
%P
)
is
i s not uniquely determined by t h e p r o p e r t i e s of 3.4.1.
It can be a l t e r e d by any element i n
H*(r
Z )[PH*(K(Z ,2n), Z p ) l . 1’ P P
High order operations
s a t i s f y i n g t h e 3.4.1 properties and put
Choose an a r b i t r a r y y
We would l i k e t o show t h a t under c e r t a i n conditions, i f
= ( x , El, y ) .
ci
X,
u E PH2n ( X , Z p ) ,
i s an H-space,
Given an H-space
!Jw[(X,Y), ( ).(
)
X,p
and w: X
X x K ( Z ,m) P
multiplication on
up = 0
then
a(u)
consists of
(hence 0 @ a ( u )1.
u
1-implications of
where
101
X
A
d K ( Z ,m) l e t
pw
P
be a
given by
(X,Y)1 = !J ( X , X ) ,
W(X,X)*Y.Y
stands f o r t h e abelian group multiplication i n
One can see t h a t any multiplication on
X
x
K ( Z ,m) P
K(Zp,m).
s o t h a t both
K ( Z ,m) + X and i2: K(Zp,m) X x K ( Z ,m) a r e H-maps must be P P of t h i s form. Indeed, i f such i s given, p1p = ppl implies
pl:
X
x
-+
N
For dimension reasons N
SI*(p2!J, Z p ) I m
=
Cl(l
+ c
2;
B lm B 1 0 1) 0 1 0
2;
+ c*(1 0
1 B 1 Q Im)
+
0 1
i As
i
2
i s an H-map,
c1 = c
P,~[(x,Y), satisfies N
p((*,*), N
!J
= lJw.
H*(w,Zp)im ( x , + ) ) = x,*,
2
= 1 and
(X,~)I = %xyY)-y.7;
= Z z ! 0 z;. i 1
Finally,
j:
x
x
x
+ K ( Z ,m) P
N
p ( ( x , + ) , (*,*))
consequently G(x,*) =
*
= W(*,x),
W
= x,* =
WA
and and
102
The cohomology of H-spaces
For any map u: X ? K ( Z ,m) P hu(x,y) = x,u(x).y.
given by
let
hU: X
K ( Z ,m) P
x
-+
X x K ( Z ,m) P
hU a r e homeomorphisms h -1 = hi;'
Then
U
and any s e l f homotopy equivalence hi2
-
h
i2 i s homotopic t o one of t h e
X x K ( Z ,m) P
of
h
isan
i s an
i
i2
2
-
i s an
i2
-
i2
p1
-
p1
-p
- p1 - u
multiplication.
essentially different
i2
- p1
-u
-
p1
equivalences and
m u l t i p l i c a t i o n and
equivalence with homotopy inverse
p1
plh
u
N
p
with
h 's.
W e refer t o such homotopy equivalences a s
it can be e a s i l y seeh t h a t i f
be
g
then
hz(g x g)
Therefore, one can ttilk about multiplications.
One has t h e
following lemma:
3.4.2.
Lemma:
The set of essentially d i f f e r e n t i2 - p1
multiplications for H ~ ( XA X,
X x K ( Z ,m) is i n
P
zP ) / i m J;I.
Proof: -
Let
w: X
A
X
-
-p
1-1 correspondence with
u: X
K ( Z ,m), P
-+
K(Z ,m). P
Consider
As H*(HD(u,u,m) , Z ) =
F*(u)
i t follows t h a t
P
and one has a s u r j e c t i o n of
p ( X
A
X, Z ) / i m
P
*;
huuw(h -1) = P u u F*(U)+W onto t h e s e t of
High o r d e r operations
essentially different y i e l d equivalent f o r some
3.4.3.
u: X
Lemma:
structure where
*;:
induced by
pw
i2
- p1
i2
+
K ( Z .m)
P
kt
X,p
-
on X
E*(x,z~)
x
- p1
-p
- 1-1
multiplications,
and w
be
multiplications.
-
1
= p*u
w: X
HA.
K(Zp,m)
F*(x,zP
103
pw
1
A
X
-+
K(Z ,m)
Q
F*(x,zP
a r e homotopic o r e q u i v a l e n t l y i f
I n functional notations:
(h
-1 o u
yiekii3 an = (1 Q
;T")w
is the reduced coproduct
pw(pw x 1) if and only if t h e two maps
are homotopic.
P
if and m ~ if y (;i, 8 l)w
p.
-
- h upw
+ w0'
Proof:
p w ( l x pw)
Conversely, i f
w0,w1 -1)
u
104
The cohomology of H-spaces
and
(14 -;.)w
3.4.4.
of
X
=
(T* Q
m e essentiaZly different
Corollary: x
1)w.
K ( Z p y m ) , where
i2
-
p1
-
HA structures
p
is HA, are i n 1-1 correspondence with
X,p
H*(X,Z Cotor
(ZP'ZP)
2, a-2
Proof:
elements of
Such H-structures a r e represented i n view of 3.4.3 by in
dim m
and by 4.3.2 any two such s t r u c t u r e s a r e equivalent i f and only i f they d i f f e r by an element i n the essentially different in
-
u* Q
im
T*:
i2
1-1 correspondence with 1
-
1 0 *;
??*(X,Z )
F*(X,Z
-u
HA s t r u c t u r e s on
P
-
p1
ker(r* 8 1
-
1Q
P
) 0 F*(X,Zp). Hence,
L*)/im
-
X
K(Zpym)
x
IJ*. But
-
p*
are t h e f i r s t two derivations i n t h e co-bar H*(X,Z
construction determining
3.4.5.
Proposition:
H*(X,Z P
Let p be an odd prime,
cocomtative.
rf
consists of 1-impZicatiuns o f
Proof:
(ZP'ZP).
Cotor
2m
x E PH
X,p
(x,zP), 2
=
an HA space with
o then
x.
Consider t h e diagram of H-spaces and H-maps:
a(x)
are and
High order operations
X: X -+ El
105
an arbitrary lifting of x.
Now 5
as an
e(6x) = n(c)e(x)
(m)
loop map is an HA map.
+ e ( c ) ( xA x
A
x).
By 2.5.2
e(5) = o
AS
and
n(5)
-
w
e(o)
=
2mp-1) with the (2.1.1) induced multiplication is an HA P' space. This multiplication being an i2 - p1 p multiplication is of and
X
x
K(Z
-
the form pw,
w
classified by
H*(X,Z Cotor (ZP'ZP) 2, 2mp-3
As H*(X,Z ) P
-
is co-associative (X,p being HA) and cocommutative
PH*(X,Z 1 o PH*(X,Z 1 P
P
H*(X,Z Cotor P 2,*
(z,z)
-
dimensions, hence, w = Cu! O u; + U*d, follows that HD(X) = jlw
-
p*H* X,Zp)y
P P
is an isomorphism in odd
ui, uy E PH*(X,Z ) .
P
It easily
and therefore
-
[H*(X,Z ) 8 H*(X,Z )](Add)*y P P
= Bw.
o
106
The cohomology of H-spaces
Let
zE H
2n
(X,Z ) P
be any element. 2%
= <(;*)'
Then,
H*(X ,ZP)y,
z e
Z Q . . . ~ P
? P
3.4.5 now follows from t h e f a c t t h a t a ( x ) = {H*(x,Z )yIX: X P
3.4.6.
Example:
I n [Harper]
For dimension reasons
dim 2p
2
+
2p.
x
3
i
an H-space
X
-+ E1,rlX
=
XI.
i s constructed so t h a t
are primitive and t h e r e i s no generator i n
3.4.5 implies t h a t
X
cannot be HA.
Another example of an application of t h i s technique i s t h e determination of t h e decomposability o f t h e Steenrod operation
3.4.7.
Proposition:
+1,
80
$2
There exist z
P
t hat
F2t11
+
B$2
-- 9
- zP
p,
p-odd.
st u b t e secondary operations
107
High order operations
Proof: c
CB = 0 = BA.
fA: K(Zp,m)
-
Let
fc: K ( Z p ,
N
it,
m+4p-3)
fCfB
N
m
K(Z
P'
fB: K ( Z p , m + l ) x K ( Z
Then: fBfA
A(p):
(Pp-2' 6)
=
Then
Consider t h e following matrices with e n t r i e s i n
x
1.
Pa
be l a r g e and
m+l)
x
K ( Z p . m + 2p
-
2)
m+2p-2) -+ K ( Z p , e 4 p - 3 ) x K ( Z p y m+2p(p-l))
K ( Z p y m+2p(p-l))
-+
K ( Z p , m+2p(p-l) + 1)
Consider t h e following:
The cohomology of H-spaces
108
f
C
Define :
If
x
E H m ( X y Z ) i s i n domain
PP-'$,(x)
As two l i f t i n g s RscgBAj = Q f c R f B element
P
+
x1
$,
n
domain $,
then
B$2(x) = { H * ( R f C ~ A XZpy11 m+2P(P-l)
-*
and
of
X
t h e set
x
w i l l satis*
Pp-2+,(x)
x1
IX:
X+ E
= jv + X
1'
rx =
XI.
and as
+ B$2(x) c o n s i s t s of the s i n g l e
H*( gcBGAxyZ ) I P e2P(P-l)*
H* ( g m ' ~3 ~'p ) 'm+2p ( p-i Z -generated by P
P p t m y hence,
E f12P(P-1)(K(Zp,m),
H*(g~BXB~' p ' lm+2p (p-1)
The proof w i l l be completed i f one shows t h a t s u f f i c e s t o determine
$i
ZP )
which i s = *PIm,
c = 1. For t h i s it
on any one c l a s s of any chosen space.
c E
z
P
.
High order operations
Fix an a r b i t r a r y fundamental class.
n
Then
and l e t iZn
I
2n E domain
E H2n(K(Zp32n), Z ) be t h e
0, n
domain $2.
P W e evaluate
Consider the diagram
P$"
AS
E
E.
1;
:I
= (n+l)pn+',
pn+l 1 2n+l
= 0
and t h e diagram i s commutative.
Looping t h e above diagram and applying 3.3.9 one obtains a diagram of H-spaces and maps:
J. K(Zp,2np)
J. K(Zp92n)
g
i s an H-map with respect t o a multiplication
pV
on
K(Zp'2n) x K(Zp92np) and t h e loop multiplication on uv[(x,y), v =
C
k=l
(&?)I
-
= x'x, v ( x , x ) - y - y ,
-1 ( pk ) ~k 2 n0
igk.Put
v: K(Zp,2n)
X = gil:
A
K(Zp,2n)
K(Zp,2n) +El.
-f
K(Zps2np),
The cohomology of H-spaces
110
HD(il,Add,pv) K(Zp,2n),
= -i v, 2
p 1 = Add
HD(X,Add,Add) = -jlilv.
hence on
El. Then, as yi
-
S i m i l a r l y , -p*H*(X,Z
P
)y
2
= H*(QfBilv,
Let
p = Add
on
are p r i m i t i v e
Z )(1 Q 1
P
2np+2p(p-l)-l
1
= 0.
Now one checks t h a t
It follows t h a t t h e r e e x i s t primitives that
H*(X,Zp)Y1
To e v a l u a t e
= -do
+
z1 and
"1 E
$l(lL)
z2 consider
zl, z 2 E H * ( K ( Z p , a ) ~ Z
P
so
High o r d e r operations
K(Zp,
2,:
H*(K(Z 211-1)’Z ) P’ P
AS
that
u z l = nj,(zP 2uGl +
tssume uz
i
= 0.
X
by
K ( ZP’ 2np+2p-4)
u3. - uLi
,2n), Z
+
di
P
a r e indecomposable primitives it follows
uzi 1
(-2p t3
+ flP1)uc2)
jl(ilx
i f necessary one may
^U2)AK(zp,2n)
As [ k e r u : PH*(K(Z ,en), Z ) P
.n dim zi one obtains
x
P
decomposable and as
md i f one a l t e r s
K ( ZP’ 2np-1)
----)
i s generated by u H * ( K ( Z
-
di
2n-1)
111
P hence,
z = z2 = 0, 1
1
2n
2n
2n
-t
PH*(K(Z 2n-1), P’
2n
Zp)] = 0
2n
To evaluate t h e l a t t e r one uses t h e f a c t t h a t it equals c
one can make t h e c a l c u l a t i o n s
)dulo t h e i d e a l generated by P ’ I ~ ~ ,i > 1, and t h e r e
1-7
,P-2(P11 2n 2n 12) =
(P?2n)P
and
c =
1.
r The decomposability o f
Pp
r > 1 i s more complex and proofs can be
und i n [Liulevicius I.
Using t h e b a s i c ideas of 3 . 4 . 5 a more r e f i n e d b u t complex technique s l d s s t r o n g e r r e s u l t s ( s e e [Zabrodsky]
192,496
and [Lin]
132
1.
Following
112
The cohomology of H-spaces
i s a sample of t h e s e results:
3.4.8.
Proposition:
If
3.4.8.1.
bH*(X,Zp)
with ker 0 c i m
If
3.4.8.2.
X,p
Let
x , ~ be an H-space. p-odd then there e d s t s a homomorphism
= 0
9.( [ZabrodsQ16)
dimensional generators then Sq4n: PH4n+1(X,Z2) monornorphism.
-
is an e x t e r i o r algebra on odd
is HC and H*(X,Z2)
is a
PH8n+1(X,Z2)
( [Zabrodsky14)
I n conjunction with t h e discoveries of p i l e s of non-classical f i n i t e CW
H-spaces t h e following was proved using similar methods:
3.4.9. n
Proposition ([Zabrodsky 15):
# 2 j -1 s a t i s f y Hn(f,Z2) # 0 .
and X ( f , X )
be the pullback of
be an H-space,
Let X Let
hi: Sn
-+
-t
S"
be a map of degree A
S"
and hA. I f
f
f: X
A
is even then X ( f , h )
does not a h i t an H-structure. Most r e c e n t l y , a long-standing conjecture was resolved for p-odd.
3.4.10. p-odd.
Proposition ( [ L i n )
Then H*(QX,Z)
1:
Let
be an H-space.
X,p
H*(X,Z
P
)
finite,
i s p-torsion f r e e .
Using K-theory operations, J . Hubbuck proved many cohomology s t r u c t u r e theorems f o r H-spaces.
(See [ H ~ b b u c k ] ~ - ~ One ) of t h e most
s t r i k i n g of h i s discoveries i n t h i s context i s :
3.4.11. Proposition ([Hubbuckh):
Let
X
Then X is homotopy equivalent t o a torus
be f i n i t e dimensional HC space. T" = ( S l ) " .
113
Chapter I V
Mod p theory of H-spaces
Introduction The source of any type of mod p theory i n a l g e b r a i c topology i s S e r r e ' s study of c l a s s e s of abelian groups, where an e s s e n t i a l meaning was given t o maps inducing isomorphisms modulo a given c l a s s of a b e l i a n groups (e.g. :
t h e c l a s s of
#
p-torsion groups).
The a n a l y s i s of S e r r e ' s observations branches i n t o two r e l a t e d b u t d i f f e r e n t methods: [AdamsI2 i g n i t e d t h e i d e a s l e a d i n g t o t h e l o c a l i z a t i o n theory formulated by Sullivan and developed t o become a s u b f i e l d i n a l g e b r a i c topology.
(We s h a l l use [Hilton, Mislin, Roitberg] as a
standard reference t o l o c a l i z a t i o n t h e o r y ) . The o t h e r method i n mod p theory i s t h e one c l o s e r t o t h e o r i g i n a l work of S e r r e : [Mimura, Oneill,Toda] and [Mimura, Toda]
1,2
developed t h e
theory of p-equivalences and p-universal spaces. The main d i f f e r e n c e i n t h e two t h e o r i e s i s t h a t t h e f i r s t c a r r i e s us o u t s i d e t h e category of (homotopy equivalents o f ) CW complexes of f i n i t e type which we t r y t o avoid, on t h e o t h e r hand one has t o concede t h a t t h e mod p equivalence theory i s e f f e c t i v e only f o r a l i m i t e d family of spaces ( t h o s e c a l l e d p-universal) while l o c a l i z a t i o n theory i s u n i v e r s a l l y effective
.
Fortunately, H-spaces a r e p-universal f o r every
p
and f o r t h e m mod p
equivalence theory i s equivalent t o l o c a l i z a t i o n theory and i t i s t h e f i r s t one t h a t w e use here ( t h e only exception i s t h e proof of 4.2.5 where we
refer t o a l o c a l i z a t i o n theory proof i n order t o avoid t h e unnecessary
114
Mod p
theory of H-spaces
development of i t s p-equivalence
equivalent).
This chapter contains a survey of t h e b a s i c f a c t s i n t h e [Mimura, Oneill, Toda] theory ( s e c t i o n 1) as w e l l as t h e notions of mod p homotopy due t o Rector ( s e c t i o n 2 ) . The decomposition of
0
equivalences of s e c t i o n 3 and t h e mixing
technique of s e c t i o n 7 follow [Zabrodsky] t h e non c l a s s i c a l H-spaces i n s e c t i o n
397
and l e a d t o t h e study of
8.
Mod p H-spaces a r e studied i n section 5 with t h e main results being 4.5.2 and 4.5.3.
The first w a s o f t e n assumed t o be t r u e though no
e x p l i c i t proof w a s produced as far as we know.
4.5.3 closes completely
t h e uneasy gap i n mixing H-spaces with nonconsistent r a t i o n a l Hopf algebras (and see remark 4.5.3.1). I n s e c t i o n 6 t h e notion of t h e genus of an H-space i s s t u d i e d and as an i l l u s t r a t i o n t h e genus of
4.1.
i s given a lower bound.
p-equivalence and p-universal spaces
Let
4.1.1.
P denote the set of primes.
Definition:
Let
p E P U (0).
p-equivalence if H * ( f , Z p ) , H,(f,Z) for
SJ(n)
Q Zp
, r(f) Q
Zp,
f: X
(equivalently
n,(f,Z
P
*
Y
i s said t o be a
H,(f,Zp),
H*(f,Z)
) ) a r e isomorphisms.
Here
Q Z
Z P
P’ = Z/pZ
p E P, Zo = 8. f
i s said t o be a
p-equivalence f o r every Note t h a t nm(fiber f )
f: X
*Y
P1 equivalence,
P1 c B,
if
f
is a
p E P1.
i s a p-equivalence,
p E P
a r e all t o r s i o n groups (of order prime t o
i f and only i f
p)
and hence
p-equivalence and p-universal spaces
every
p
4.1.2.
equivalence i s a
Definition:
P1 c P U {O}, f = f
q.nO
0-equivalence.
A space
where
P1 Universal,
i s s a i d t o be
X
P1 and n 0 > 0 t h e r e exists i s a P1 equivalence and H n ( f , Z ) = 0 f o r
i f f o r every
: X + X
115
q
f
9
0 < n zno.
4.1.3.
Definition:
A space
i s s a i d t o satisfy a finitenass condition
X
if (f.c.):
Hn(X,Z) = 0
(f-c.H)
n > N(X)
for
The following only s l i g h t l y d i f f e r fromtheorems 2.1 and 3.2 of
[Mimura, Oneill, Toda].
4.1.4.
Proposition:
(a) X
For a space
the f o l l m i n g are equivalent:
X
is P1 universal
( b ) For every P1-equivalence satisfying
(f.c.n)
f: X’
(f.c.n)
X
with f i b e r f
there e x i s t s a P1
( c ) There e x i s t s a P1 equivalence satisfying
-+
and X o
f: X
-+
equivalence X
0
Y,P, P1-equivalence
f i b e r satisfying
and a map
P1 equivalence H
hf
- %I.
f: X
-+
X
n0,q
:
x-+x
so that
X I .
with f i b e r f
3:X
and a map h: X
N
f: Y -+
-+
-+
with
y, there etcist a Y
so that
(el For every integer no and a prime q # P1 there e d s t s f = f
-+
is P1 universal.
( d ) For arbitrary spaces
( f .c . ~ )
g: X
an(f,z ) = 9
o for o
< n :no.
116
Mod p
If)
theory of H-spaces
For every P1 equivalence (f.c.H) there e x i s t s a
f: X +
XI
P1 equivalence
( g ) There e x i s t s a P1 equivalence f : Xo satisfy i n g
and Xo
f.c.H
is P
1
N
( h ) For any P1 equivalence
f: Y
( f . c. H) and any map h : Y N
and a map h:
f: X + X
Proof:
(a)
I ,
(b).
with cofiber s a t i s f y i n g
.+
-+
XI
X
with cofiber
-+
universal. with cofiber s a t i s f y i n g
there e x i s t a P1 equivalence
X
P .+ X
P
-+
X.
g:
-
so that
fh.
The proof c o n s i s t s of t h e following s t e p s :
Decompose
f:
XI
-+
X
i n t o i t s p r i m i t i v e Postnikov
decomposition (which i s f i n i t e as involves o n l y primes
X' =
qi
xm
Consider:
I
km
t fm
as
p1
~
xm-l
satisfies
fiber f f
i s a P1
- ..
.
(f,.c.n)
equivalence):
and
117
p-equivalence and p-universal spaces
fn f
-
t h e f i b r a t i o n induced by : X
Sll'
-+
Jn j
.
n
If
then as kngn-l?
^f=?
Sl.J1
^r
i s a B1
i < n
gi
N
R 'Jn
,Z
. . . 2G J m i s
is a
f
91Jl
exists so that
*,gn
g: X
Bel
a
s2,J2
equivalence
Then
?
e x i s t with
f
R
... f
S21J2
fng,
N
) = 0
for
flf 2 . . . f .ig i
N
(lidi
gn-lf R'Jn'
-+
equivalence,
fg*-
f
gm = g
and
X'.
m
H ( ,Z ) 9
y i e l d a s u r j e c t i o n of j.
jn-l),
Qn'
R'jn
mL
0 <
g:
f i b e r K( Z
with
P1 equivalence with ?(?
is a
X
kn
f: X'
Let
m 5 no.
P1 equivalence and f i b e r f
+
X
satisfy
be t h e f i b e r of (f.c.n).
If
W P
x
X' then
fl(f,Z
9
) = 0
f
0 < mLn
for
= gf
9,n0 0'
is a
Hm(f 9.n0
P1 equivalence and as
,Z
P
= 0
O < m c n
0'
(a) * ( c ) t r i v i a l . (c)
r,
Hm(?q
(a).
)no
,Z
Given
9
) = 0
n
0
and
q
let
5 no.
0 < m
^f
=
2
xo
(b)
*
ax.
Put N
(d).
Given
f
q 'no
By ( b ) applied t o
= gff.
N
f,h
X
form t h e i r pull-back:
h
Xo
1 > Xo
9 'no
W P g:
RIP :
N
+Y
Xo
satisfy
t h e r e exists
118
Mod p
theory of H-spaces
f i b e r ? = f i b e r flafl
As
is a P
1
equivalence and
f i b e r fl
U P satisfies
(f.c.n),
f = flgl,
h = hlgl.
(d)
*
by (b) t h e r e e x i s t s
gl: X
5
?.
Put
N
Given
(b).
f : X'
X put ? = X ,
N
Y=X',h=l,
f = f
U P and by
(a) one
has
equivalences s o i s
(a) * (el If
fl: X
x
5
ifl =
a. AS
fafl
h : X -+ XI.
follows from t h e following m r e general observation: f: X
m = m(no)
+
X
so that
Indeed, i f Hm(f,Z) €3 Z
9
satisfies p ( f , Z ) = 0
9
rn(?,Z
) = 0 9
X = K(G,n),
= 0
for
Hm(f,Z
n
0 < n < no
) = 0
9
and Tor(Hm-l(f,Z),
for
5 no then t h e r e e x i s t s
- 2.
m 5 n + l 5 no
Zq) = 0 ,
implies
therefore
and T o r ( n n ( f ) , Z ) = 0. But nn(K(G,n), Z = Tor(G,Z 1, 9 9 9 n (K(G,n), Zq) = G Q Z and nm(K(G,n), Z ) = 0 f o r m n-1 9 9 n * ( f , Z 1 = 0. 9 Suppose, inductively, f o r integer
are
m = m(n)
n < n0
-
nn(f) Q Z
# n-l,n,
9
= 0
hence
1 t h e existence of a p o s i t i v e
s o t h a t i n t h e following commutative diagram
p-equivalence and p-universal spaces
Then as
cofiber hn
is
n+l
connected and
119
( c o f i b e r h ) = 71 (X) n n+l
H n+2
one has
H j ( X,Zq)
(hn)"
I
H ( X , Z 1J n s
H ( c o f i b e r hn,Z )j 9
Hj-l(XyZq)
I
g=0
Hj(XyZq)-
H ( c o f i b e r hn,Z 1j 9
Hj(Xn,Zq)-
= 0,
and hence
j 2 no
Hj(f N2m , Zq) = 0,
H,(?
2m
Htn+2
= 0.
= 0,
Hj(32my
j ( n + 2 z n o y imply
Together with
nj((fn+l)4my
, Zq)
(XyZq)
kn+l
Htn+2
implies
J-1
and i n the following
kn+l
IT*(?^, zq)
H
z,)
= 0,
n ((f,)",
j z n + l < no.
(H*(K(G,k), Z ) 9
Steenrod algebra A ( q )
by
$1.
Z
j
Z ) = 0, 9
= 0,
9
j
j (n+2
5 n+l
and
< no,
~ , + ~-2m ( f,zq) = 0
this
implies
i s generated as an algebra over t h e
Now
f 4m
n
9
4m
fn+l'
ph,
y i e l d a comparison
Mod p
120
theory of H-spaces
of t h e Serre s p e c t r a l sequences of the f i b r a t i o n s
Hj((fn)
4m , Zq) = 0 = H j ( ( 6 ) 4 m , Z
and consequently t h e other by the 0-map. 0 = Eo[(fn+l)
4m
1:
Em
9
j <-no
implies t h a t t h e
Consequently, i n
Eo[H,(Xn+ly Z q ) l
Put
Zq), j 5 no,
m(n+l) = 4m(n).no
+
dim :no, Zq)l
Eo[H,(Xn+l,
t o conclue (e)
*
(a).
j
5 no
The proofs ( e )
( f ) e+ ( h ) are theEckmann-Hilton duals of ( a )
(*
and
4mn0 Hj(fn+l ,Z
z91
j < no-2
Zq)
t o complete t h e inductive s t e p .
n ( f ,Z ) = 0 , j q
E0 i s the
where
H,(Xn+l,
lowers f i l t r a t i o n , hence,
One can dually prove t h a t
terms
of t h e s p e c t r a l sequences are mapped t o each
grading associated with t h e Serre f i l t r a t i o n of Hj(fn+l, 4m
E2
=
imply
**
9
= 0.
Obviously
o
for
j < n0-2.
Hj(fm,Zq)
( f ) , (e)
= 0
( g ) and
( b ) , ( a ) c) ( c ) and ( b )
0)
(d)
respectively where pushouts replace pullbacks, Moore decompositions replace Postnikov approximation and cohomology and homotopy interchange.
4.1.4. 4.1.3
Corollary: StU48
If
x
is P1 uniuersal s o are H t n ( X )
v a l i d if the f i n i t e n e s s conditions imposed on
and H ~ ~ ( x ) . fiber f
in
p-equivalence and p-universal spaces
( b ) , ( c ) and
(a) and those for cofiber
in
f
121
and
( f ) , (g)
are
(h)
repkced by the conditions that a l l spaces i n 4.1.3 satisfy a f i n i t e condition of the same type Proof: -
(f.c.n
o r f.c.H).
The P1 u n i v e r s a l i t y of
IEP. ( X )
d e f i n i t i o n of universal spaces 4.1.2. follows from 4.1.3 [Htn(X), Htn(Y)]
(a). The
follows d i r e c t l y from t h e
n
P1 u n i v e r s a l i t y of H t n ( X )
The
second p a r t follows from the isomorphisms
-% [Ken(X), mn+l(Y)]
and one can move from one t o
another i n the proof of 4.1.3.
4.1.3 (a) e a s i l y implies t h a t any H-space i s P1 universal f o r all But one has a stronger result:
P1.
4.1.6.
Proposition ([Mimura, TodaL, Theorem 1 . 2 ) :
0-equivaZence. SO
is
If
+
Xo
be a
is P1 universal
and Xo
x.
Proof:
Decompose h
x = xm h j : Xj
s a t i s f i e s f.c.n
fiber h
Let h: X
-+
Xj-l
i n t o a primitive decomposition.
hm
xm-l
induced by
k
X
*
j'
j-1
+
. ..
ho
xo
, n 1.
K(Z
qJ
We s h a l l use induction t o show t h a t
9
X
j j
is
lP1
universal hence, it
s u f f i c e s t o prove: Let
h: X
-+
Xo be a K(Z
Pa
k: X + K(Z , n ) . Then, i f P Fix q # P1 and no, n equivalence with
H (f
m "0
,
> n.
z A )
9
p r i n c i p a l f i b r a t i o n induced by
i s P1 universal, s o i s
Xo
0
n-1)
=
If
o
p for
# P1 l e t
fn : Xo 0
= q and f o r
X. -+
Xo
be a P
= p a m ( n 0'
1
Mod p theory of H-spaces
122
Then one has a commutative diagram
2.
X
r
+ X
0
c
xO
xO
1.
I
A
As
is a P
h
equivalence.
equivalence,
1
+ q,
p
If
a P equivalence hence, 1
f
Hm(l,zq)
=
o
H*(h,Zq)H*(fn ,Z ) = H*(lf,Z )H*(h,Z 1, 0 9 9 9
p(fn ,Zq) =
m 5 n 0'
for
0
0
If
for
i s a P1
m 5 n0 : i s an isomorphism and
H*(h,Z ) 9
p = q
f
consider t h e S e r r e s p e c t r a l
sequence f o r hamology i n t h e diagram
-
~ ~ ( 2 . 1E :~ ( x ) 2 -
,z 1 Q
H,(X
, Zp)
E ( f ) = H,(fn
Q H,(r,Z
0 Em (1)= 0 m,n
and
H , ( K ( z ~ , n-11,
zP)
+
E 2 (XI
O P
N
-
H,(f,Z
P
)
P
)
and
E2 m,n
(1) =
0,
lowers f i l t r a t i o n i n
0 < m+n
2 no, n
dim
5 no.
Hence,
- 0
f
123
p-equivalence and p-universal spaces
n
i s a p1
Em(? 0 , zP)
equivalence and
Suppose p E El. n ‘fq,nO
, zq)
r = rank Hn(X
=
o
Let
for
o
<
f
9rn0
: Xo
=
+
o
m 5 no.
for
0 be a P1 equivalence,
X
Hm(fQInO, zP) E
m <-no.
hence t h e r e e x i s t s an i n t e g e r
Z ), 0’ P
GL(r,Zp)
a
so t h a t
n a H (fq,no, Zp) = 1 and one h a s a commutative diagram
and as
is an isomorphism,
H*(h,Z 9 )
Finally, as
.
f
H * ( f aq,nO, Z ) = 0
induces a homotopy equivalence
4
H*(?,
ZQ ) = 0.
f i b e r h -+ f i b e r h
the
upper square of t h e above diagram i s a p u l l back diagram, fiber
4.1.”.
*r
P
= fiber fa
and
Definition:
A space
q+)
P
is a
1
equivalence.
i s c a l l e d an
X
Ho
f r e e ( a s s o c i a t i v e , commutative graded) algebra.
space if H*(X,Q) (e. g:
is a
Every H-space i s
an Ho-space).
4.1.8.
Corollary:
Proof:
An Ho
There e x i s t s a
space 0
X
is P1 universal for every s e t
equivalence
X
+
K(n,[X)/torsion)
l a t t e r is P1 universal, hence, 4.1.8 follows from 4.1.6.
5.
and t h e
124
Mod p
theory of H-spaces
Mod p homotopy
4.2.
Most of the notions and properties discussed i n t h i s s e c t i o n are due F i r s t we formulate some simple properties of
t o Rector (unpublished). P1
equivalences some of which w e have already used i n t h e previous
sections. 4.2.1.
Lemma:
( a ) Consider the follaving diagram: h
Y
4.2.1.1
I f 4.2.1.1
+x
is e i t h e r a puZlback or a pushout diagram then
is a P1 equivalence if and only i f f b l I f i n 4.2.1.1
B
- P1
gl,g2
h
ho
is.
are P1 equivalences and h,ho
are
equivalences then 4.2.1.1 is both a puZZback and a
pushout diagram.
Proof: IT
k
(fiber f )
( a ) A map
f
is a
P1 equivalence i f and only i f
are f i n i t e groups of order prime t o
H (cofiber f ) k
are t o r s i o n groups of order prime t o
a pullback (pushout) diagram then
fibers (cofibers)
h
and
ho
(or i f and only i f
P1). If 4.2.1.1 i s
have homotopy equivalent
.
(b) Form t h e pullback of them t o obtain:
P1
ho.g2
(pushout of
h,gl)
and complete
Mod p
hmotopy
h
Y
1
I
f2
1;
being
U
xo
Similarly
f.
p,
f
g2
is a
f
P
fl
-
and as
' xo
Being a
P1 equivalence.
i s a homotopy equivalence.
P1
is a
g1
Similarly
f.
4.2.2. ape
gl
P1 equivalence implies t h a t s o i s
p-equivalence f o r every for
f; V
hO
equivalence s o i s
X
I
___+
g2
k
n
'X
"ho 9 g2
125
Lemma ( R e c t o r ) :
kt
X
fl,f2:
+ Y be maps.
Then the following
equivalent: ( a ) There e x i s t s a P
1
( b ) There e x i s t s a P1
Proof:
(a) * (b)
.
al'a2
-x
equivalence
g : X1
-+
X with flg
equivalence
gl:
Y
+
Y
For any maps
>
Y j
b e an e q u a l i z e r diagram (see e.g.
al,a2:
1
X +Y
with
-
glfl
f g. 2
-
glf2.
let
'E a13a2 [Spanier], p . 406).
Consider t h e diagram
Mod p
126
theory of H-spaces
J,
Y
The cofiber of
^g i s of t h e homotopy type of t h e suspension of t h e
cofiber of
Hence,
g.
g
is a
P1 equivslence.
t o the existence of a l e f t homotopy inverse f o r
flg j,,
N
f 2 g is equivalent
hl: E
flgaf2g
+ Y.
Consider
Y
h.
J,
Y
IA
I"
Ig1
with the r i g h t hand square being a pushout diagram.
Then
g1
is a
P1
equivalence and
( a ) i s dual (using cc-equalizers).
(b)
4.2.3.
Definition (Rector)-:
(fl
f2)
flg
N
f2g
equivalence
flaf2: X -+ Y
we s a i d t o be El homotopic
i f one of t h e equivalent conditions of 4.2.2 are s a t i s f i e d :
f o r a P1 equivalence gl'
g
or
glfl
N
glf2
for
B1
Mod p
4.2.4.
-
Lemma:
homotopy
127
i s an equivalence relation respecting composition.
i s obviously r e f l e x i v e and symmetric. Let g2: Y .
.
-+
Y
A
*gi
and as
2
are
g2g2
x1
f;,,fi:
3
Y,
f o -P1 fl,
equivalences,
p1
fl
glfo
N
x2
i s a P1 fo
NP1
equivalence f;,
f l y
f o g - flg.
Y2,
fbfl -pl f i f l
wP1
gl: Y
Let
f2.
g2fl
-
Let
fo,fl:
glfl,
g2f2.
-+
Y
1'
Let
gl,g2
equivalences.
x2,
equivalence
gl:
P1
be
+
complete a pushout diagram f o r
gl,g2,Yo
then
X
fo,fl,f2:
Then
glf;
-
Npl
fo
fi.
fbfog
glfi.
wP1
N
Xo
+
5,
n
g: Xo
Let fhflg
Then
and by t r a n s i t i v i t y
f2.
-+
Xo
be a
P1 Let
and
g f'f 1 0 1
-
g f'f 1 1 1 and
fAfo -pl f i f l .
Using [Hilton, Mislin, Roitberg] Theorem I1 5 . 3 and t h e f a c t t h a t a
P1 equivalence P l o c a l i z e s t o 1-
4.2.5.
Proposition:
or i f X s a t i s f i e s every p
implies
Given spaces ( f . c . r ) and
fl
a homotopy equivalence ane has:
-
f2.
Y and X . fl,f2:
Y
-+
If Y
x
satisfies ( f . c . ~ . )
then fl
Np
f 2 for
128
Mod p
4.3.
Decomposition o f
0
theory of H-spaces
equivalences
We discuss in this section ideas introduced in [Zabrodsky]
337‘
4.3.1.
Proposition:
P1
$I:
$: X
+
Xo
be a
there e x i s t s a space
o f primes Ply and maps
Let
X
+
X(P,,$),
equivatence and
$I1
+
Given a s e t
unique up t o homotopy type
X(Pl,$)
X(Pl,$)
$I1:
0-equivalence.
x0
s o that
-
is a
$ “ J , ~ $,
is a P
- P1
- a set
o f primes, then there e x i s t s a map
equivalence.
Furthermore, given a
commutative diagram
$,T
0-equivalences,
P1
f ( P 1 y $ y ~ )so y t h a t the following i s commutative:
If both X and Xo f(Pl,$yT)
(f,c.H)
or
xy?o s a t i s f y
(f.c.s)
then
i s unique t ~ ,t o homotopy.
Proof:
x(P,,$),
and
$I1
$A: X + Xn’ $:: 0-equivalences, A$:$ - I+,
Suppose are
satisfy
Xny
are constructed inductively as f o l l o w s :
Xn + X
is a
0
are given so that,
P-P1
$A,$:
equivalence and for
Decomposition of
equivalences
0
p E P1 Hm($AyZp) i s an isomorphism f o r m = n,
or equivalently
prime t o Let
H (cofiber
m
m < n
$A, Z )
129
and an epimorphism f o r
are t o r s i o n groups of o r d e r
P1 f o r m ( n . Vn = c o f i b e r
Consider t h e compositions
$ I .
n
kn+l
P i s t h e sum o f t h e p-primary components of
bn+l(Vn)
Hn+l( Vn , Z ) = Hn + l ( Vn 1, Note t h a t as
i s a 0-equivalence
$'
n
natural.
Obviously
to
x
obviously a
Finally,
-+
xn+l'
kn+l$A Put
Y
$A+1y
z)
i s t o r s i o n and
Hn+l ( V n y Z )
and i f
X = fiber k n+l n+l
= $:hn+l,ns
hn+l,n: Xn+l
P-P1 equivalence, hence so i s
Z) = H (cofiber m
+
X(P1,$)
t h e composition
X(Pl,$)
Lemma:
lifts
Xn* hn+l,n
$A,
Z)
for m 5 n
is
and
= l i m Xn, an
= l i m $I + n
$I
f
-
which completes t h e and
$'I:
X(Pl,$)
f(Ply$,;)
-+ Xo
-
$;
' n'
Xo
(which i s independent of
The uniqueness of homotopy type w i l l
follow from t h e second p a r t of t h e proposition.
4.3.1.1.
is
n+l
$:+l.
b n + l ( c o f i b e r $'n' Z )
a r e t h e d e s i r e d space and maps.
The e x i s t e n c e of
$1:
k
P-P
=
inductive s t e p .
n)
N
Hm(cofiber
Hn+l(cofiber
P E P1.
follows from t h e following
civen a conunutative diagram
Mod p
130
There exists ?: Y1
-+
Proof of 4.3.1.1:
80
theory of H-spaces
that
?g
-
fo, h?:
-
Construct the pull back
fl.
W
of
fl ,h
fl
and
h
and complete:
As
;i
is a P-P1
f is a
?'
(a)) and
g
0-equivalence and by the first part of 4.3.1
exist.
are P
equivalence (by 4.2.1
2."
and
are P-P1
equivalences, hence, so is 1-
equivalence.
One has
is a P1
Y(P,,?),
equivalence and so is n+.
gf"
and
3'
*c
gf".
is a homotopy
equivalence
?*
and and
g
Decomposition o f
equivalences
0
fO
N
ga”
N
M
ry
f =
gf”).
y1
lfl
(Simply choose
1.
?:1a”
a” = ?“u
$1,
h =
-, $11
-
fo = $ I f ,
Y = X,
?:
Finally,
u
i s t h e homotopy inverse of
N
?g = fla”g
N
Y1 = X(PlY$),
fi = f0$”
uniqueness of homotopy type of
and
where
i s t h e desired map:
Now, apply 4.3.1.1 f o r g =
131
X(pl,JI)
t o obtain
fo,
H = y(Pl,T),
-
X1 = Xo,
The
f(Pl,$,T).
follows from 4.3.1.1:
is a P
1 and a P-P1 equivalence, hence a homotopy equivalence. f(P,,$,v) i s unique i n t h e f.c. case by 4.2.5: If given
X
I
JI’
’ xo
I
I
fo
Mod p
132
pTl
- ?C?,
implies
N
fl
P-P1 f 2
-
N
and
f1$'
y1z2, ( n o t e t h a t by uniqueness
and by 4.2.5
X(Pl,$)
N
theory of H-spaces
-
X(Pl,T)
and
they i n h e r i t t h e
N
N
f2$I
implies
N
f l-P1 f 2
o f t h e homotopy type of
f.c.
of
X,Xo
and,
X
M O
Xo
--
X,Xo
respectively).
4.3.2. then
Corollary: If X,Xo
X(Pl,$),
and
$I
$I1
are H-spaces,
$:
is an H-map
acbrrit H-structures.
Proof:
By t h e uniqueness o f t h e 4.3.1 decomposition one has
Y)(P,,$
X
11.1
(X
X
$
an H-map and a
CJ
X(P1,$) x ?(P,,T)
and i f
X,v,
Xo,uo
a r e H-spaces,
0-equivalence one has a commutative diagram
Decomposition of
4.3.3.
Let
Lemma:
.
p a r t i t i o n of P Proof: -
Jl::
X(Pj,$)
$;
and
JI:
x
- 0
0
and l e t
Xo
{€!,I
i = 1,.
Then X i s t he pullback o f t he maps
We prove t h a t t h e pullback of
*
Xo
X(Pi U P j y
is
$1.
133
equivalences
$I!:
1
-+
be a
x(Pi,$)
$;:
X(Pi,$)
..,k)
Xo
-t
x0.
and
Indeed, consider t h e pullback o f
and complete
J
$3
-
J
As
$:
is a
P-P j N
equivalence.
$!'
equivalent so i s
= $;hj
is a
(&Pi)
h
P-P
J'
n
xO j
S i m i l a r l y hi
(P-PJ)
is a
P-Pi
equivalence while as
1J
P-P a
3
Pi
is a
h
j
j
Pi
equivalence.
and by 4.3.1
Pi
equivalence and s o i s
Similarly
ij
is a
P
j
$.:
Consequently
7'
equivalence, one g e t s
id
is
Mod p
134
4.4.
A Study of
theory of H-spaces
H -Spaces. 0
Recall t h a t an
Ho
space i s a space
cohomology algebra.
If
G n
W
QH"(X,Z)/torsion and a n : Gn
a l e f t inverse of t h e projection by
JIn: X - + K(Gn,n).
JI: X
X
If
= nK(G,,n)
+ K(G,)
an isomorphism
i s an
-+
Gn
then
n*(X)/torsion. =
K(C,)
n
n
JI,
is a
Hence,
K(Gn,n)
an
-+
Hn(X,Z)
is
can be r e a l i z e d
space then
Ho
0-equivalence and induces
QH*(JI,Z)/torsion. Note t h a t f o r an
QH*( X,Z)/torsion
i n the form of
Hn(X,Z)
induced by
n
with a f r e e r a t i o n a l
X
where
space
X
spaces have r a t i o n a l models
Ho G,
Ho
= {Gn} i s a f r e e graded
abelian group.
Here a r e some simple observations and notations regarding H
K(G,)
and
spaces. 0There e x i s t s a n a t u r a l monomorphism
+( G,)
Furthermore, i f r i n g generated by
where
i s t h e f r e e ( a s s o c i a t i v e commutative graded)
t h e r e e x i s t s an i n j e c t i o n
G,
stands f o r t h e set of r i n g homomorphisms (one has an obvious
HGm
4: Hom(G,,Gi)
injection
For any
Ho
space
QH*(X,Z)/torsion X
-+
X
H;m($(G,), and
H*(x,z)
$(G&)).
X: QH*(X,Z)/torsion - + H * ( X , Z )
projQH*(X,Z)/torsion
-
so t h a t
is a
monomorphism of m a x i m a l rank t h e r e exists
JI = $(XI: X such a
X
- 0
K(QH*(X,Z)/torsion) r e a l i z i n g
a rational splitting.
Such
X
X.
We s h a l l c a l l
induces a morphism
A study of Ho-spaces
i:
4(&H*(X,Z)/torsion)
H*(X,Z).
-+
Let f: X
-+
135
XI be a map between Ho
spaces and given rational splittings: X: &H*(X,Z)/torsion
-+
X: &H*(X' ,Z)/torsion
-
For a vector X = (A1, A2,
...)
= (Al,- ..,An,. ..)
. a : 0(&H*
('XI
-+
H*(x' ,z)
of integers let I-€ Hom0(G,,G,)
endomorphism (I-)n = Anal: G n x
-
H*(X,Z)
be the
h -+
Gn. Then one can alwws choose a vector
of integers and a morphism
,Z)/torsion)
+ $ (&H* (
x,z 1/torsion)
The geometric realization of this observation is given by:
4.4.2. Lemma: ( a ) Let X be an €Io-space. QH*(X,Z)/torsion
-+
The rational s p l i t t i n g s
H*(X,Z)
are in 1-1 correspondence with
the rational equivazences $: X ( b ) FOP any map $: $I
x -+xo:
-X =
X'
-+
f: X
-+
X'
-+
Xo = K(&H*(X,Z)/torsion).
and rational equivalences
K(QH*(X,Z)/torsion),
Xb = K(&H*(X' ,Z)/torsion)
(Al, A2,.
.. ,A,,..
.)
there exLsts a vector
of integers and a map
fo: Xo
-+
X' 0
(QH*(fo,Z)/torsion = w*(f ,Z)/torsion) so that
4.4.3. $:
X
-+
Corollary:. Let
X,u
Xo = K(QH*(X,Z)/torsion)
be an H-space.
Suppose
is such that QH*($,Z)/torsion
is an
Mod p
136
isomorphism.
-A =
theory of H-spaces
Then there e x i s t a multiplication m for
(a 1, X2,. . . )
of integers so t h a t
and a vector
Xo
is an H-map.
I'(I-)$ A
Apply 4.4.2 for $ x $ : X x X
Proof: p:
X
x
X
-+
m:
X. Then t h e r e e x i s t s
X
x
-+
X
X
x
-+
Xo
0
Xo,
$:
and
X
x
-+
X
and
0
so that
Consider t h e following:
x0
xO
xo
x
xO
U
imply
W*(Z
E'
Z)/torsion = 1
+nd consequently
i s a homotopy equivalence with a homotopy i n v e r s e
an H-structure for
Xo
and
r(I-)$
is a
YE *
H-map.
p-m
A
4.5.
4.5.1.
Mod P1
Definition:
x
satisfies, i : X
H-spaces
*X
x
-+
X
equivalences )
A mod P
p(x,*)
and
1 H-space i s a p a i r
x
-+
p(+,x)
are the i n j e c t i o n s ,
.
E
are
= 1,2
X,p
where
p: X x
P1 equivalences. then
h
= piE
are
X
-+
X
(i.e. i f
5
Mod P1 H-spaces
4.5.2.
Let
Proposition:
satisfy
X
mod P1
H-space
and P1
is
X
satisfies
h: X
as w e l l ) .
-
X' g ' ( h
Now suppose t h a t
hE
are
* X. Put P1
h
X
an H-space o r i f Hence,
i s an
X
is a
X
KO
space
X
h)
is a
Let
mod P iE: X
X.
mod P
-t
$: X
mod P
H-space with a
X
-f
x
X
E
= 1,2
-
i s an
pl
H-structure f o r
1
1
If
X.
H-structure
1
the injections.
Xo = K(r,(X)/torsion)
be any
Using t h e methods of 4.4.2 we s h a l l show t h a t t h e r e e x i s t
-
A = (A1,
A2,.
..)
and an H-structure
m: Xo
x
X
0
* X
0
so that
Again, t h i s i s a geometric r e a l i z a t i o n o f an a l g e b r a i c observation: G,
Put
= &H*(X,Z)/torsion, a : H*(X,Z)
and a morphism
X'
W P g: X'
and
is a
(and we m a y assume
X'
4.1.4 ( b ) implies t h e existence of
X'
= piE,
equivalences.
0-equivalence.
a vector
' -
P1 equivalent t o an H-space
H-structure f o r
X
X'
i s a Hopf algebra.
H*(X,Q)
(f.c.r)
equivalences
X
X' with
universal.
If
p: X
X-
If e i t h e r
is a mod P1 H-space
X
(f.c.a).
to an H-space.
if and onZy if X is P1-equivaZent Proof: -
137
8 H*(X,Z)
m*
X: 4 (G,) +H*(X
+
H*(X,Z)
x
X,Z).
t h e morphisms induced by One i s looking f o r a v e c t o r
so t h a t t h e following diagram w i l l commute:
Let
9.
A
138
hr
h* 1 L
H*(X,Z)
8 H*(X,Z)
a
This can be seen by induction:
Let
G
-
were found s o t h a t t h e diagram commutes with
in: $(G,,)if
-p*
+H*(X,Z)
= p*
replacing
- p*h* - p*h* 11 2 2
G,,
-
A , m*
k e r i* @ 1; = im[H*(X 1
im[Hn+l(X
+
A
X,Q)
Hn+'(X
x X,Q)I
A
X
X, Z]
respectively.
F*(X x
X, Z ) ] .
n
n 1,
m;,
Now,
*Xn+l: Gn+l + H"+l(X +
m*
hl,...,hn,
n = ( A l,...,X
-
and
Suppose
-X
G
then t h e image of
i s contained i n
Gm.
@
mzn
x
x,
2)
As
c {H*(hl,Q)(i Q Q)($(I-) Q Q)[rP(G
-
x
11 s o that
Mod P1 H-spaces
Let
$(G
-
-+
= rl Q 1 + 1 Q Now t h a t
O ( G < , + ~ ) 0 $ I ( G < ~ + ~ ) be induced by
-
-
+
mn+l 7
m and
(9: 2
+
139
m* n
and by
the unit).
.#(G,)
were defined one g e t s a geometric r e a l i z a t i o n
IJ
x x x
'X
J,
Put
l'(I-)$ = $-.
x
X(Pl, J,-)
x
x
N
X(E',,
As hg, J,-h
x E
)
E
= 1,2
are
P1 equivalences
and by 4.3.1 one obtains a commutative diagram
Mod p
140
theory of H-spaces
,x
X
x
x
The uniqueness of 4.3.1 and t h e r e l a t i o n s
imply
= 1, u(P1)
p(Bl)iE
4 . 5 . 2 . 1 . : We make proof o f 4.5.2
t h e vector
X
-5
X(B1.
$ 1 X
no minimality claims for the vector
-X
i n the
and one can apply t h e same method of construction i f one i s
-
given a f i n i t e number of and a fixed
i s an H-structure and
JI: X
-X
mod B1
N O
Xo.
H-structures
mi
for
X ( i = 1,
...k)
If such a family i s given one can choose
simultaneously f o r a l l
e x i s t s an H-structure
ui
i
s o t h a t f o r every
i
s o t h a t the following i s commutative:
there
141
Mod P1 H-spaces
x x x
x x
x0
x
"i
x
1
xo
xO
I
I
-1 x0 4.5.3. only if
Proposition:
x
xo
Let
m
i
X satisfy
xo X is an H-space if and
(f.c.).
X is a mod p H-space f o r every prime
4.5.3.1.
Remark:
cw complexes.
p.
The e a r l y versions of t h i s statement considered f i n i t e
(See [ ~ i s l i n ] ~ Theorem , 2.1).
The proofs rely upon a
preliminary announcement of Curjel (see [Curtis] which was l a t e r r e t r a c t e d ( [CurtisI1, p. 1120).
1'
remark ( e ) , p. 8-9)
The corrected versions of
the proof ( e . g . : [Hilton, Mislin, Roitberg], Theorem I11 1.8) had t o assume some r a t i o n a l consistency condition of the various H-structures of
mod p
X. This r a t i o n a l r e s t r i c t i v e assumption i s removed i n
our proof. A s t o C u r j e l ' s announcement t h a t s t a t e d t h a t every H-space with
f i n i t e dimensional r a t i o n a l cohomoloey admits a multiplication inducing a co-associative (hence primitevly generated) r a t i o n a l cohomolo@;yHopf' algebra.
A s far as we know t h e v a l i d i t y of t h i s statement f o r f i n i t e
Mod p
142
theory of H spaces
However, i f one considers
dimensional H-space i s s t i l l an open question. ( f .c.a)
H-spaces s a t i s f y i n g
and having f i n i t e dimensional r a t i o n a l (See Stasheff ' 6 H-spaces problem 44,
cohomology, t h i s statement i s f a l s e .
Neuchatel, Springer Lecture Notes 196, p. 134).
This can be seen by
observing the following example. Hta(SU)
Let
v: Ht21(SU)
Ht
h
x
(SU) 21
K(Zy23)
-+
H*(v,Z)i = u Q u u
5 13
5
u
a generator of
i
be the
- 'b
PHi(Ht2,(SU),
mod 3 reduction of
1 = p3x*3
- Y5Y9Y13
@
"5"9
2) W Z .
is primitive and i f
for
= H*(SU(12), Q ) .
The generator i n
&H23( X , Q )
i:
3
23
t h e image of
-
u5%J
"9
@
Then
i s t h e f i b e r of
X
i s an H-space,
nn(X) = 0
cannot be represented by a primitive
i s represented by a primitive, as
E H23 (X,Z )
u5
u.1 Q 1 respectively.
then
regardless of t h e H-structure chosen f o r
23
u5'13
Let
X
K(Z3,27)],
and H*(X,Q)
+
and o f
10 1
z E [Ht21(SU) x K(Z,23),
n > 26
uv,
K(Z,23) be given a twisted multiplication
x
23
X.
H*(X,Z)
Indeed, i f t h i s generator is torsion free i n
dim < 23,
i s primitive modulo a decomposable
23 + ;1 E PH ( x , z ~ ) . This implies , . . - . A
x23
+ pl; = y y y
5 9 13
+ P 1 i E PH27(X,23).
As there are no decomposable primitives of odd dimension i n a Hopf algebra 1-
;5$9i13 +P d hold i n
= 0.
One can e a s i l y see t h a t t h i s r e l a t i o n must already
H*(Ht21(SU), Z3)
(or even i n
H*(SU(11), Z 3 ) ) .
But
143
Mod P1 H-spaces
As (N
3’
X
+ wi
o
g imp1
w5*w9’w13
satisfies
E P H ~ ( S U ( U ) , z3).
i s an H-space i f and only i f
(f.c.H)
HtN(X)
s u f f i c i e n t l y l a r g e ) i s an H-space s u f f i c e s t o prove 4.5.3 f o r
satisfying
(f.c.n).
4.5.3.2
X
c a r r i e s t h e main technical. burden of t h e
proof of 4.5.3 and e s s e n t i a l l y shows t h a t t h e r a t i o n a l consistency ( f .c.n
condition always holds f o r H-spaces s a t i s f y i n g
4.5.3.2.
satisfy
equivaZences,
H-structures
N
pi
yym2
so t h a t
no > n
N
i ,Z P 1 =
Proof:
n
#.
Ti
is a
and
P2 = pi
$i:
be H-spaces,
xi + x o
Suppose
- mi
H-map.
y2 respectively so
Xo
80
Xi
i = 1,2
a&ts
Then Xi
and
t h a t $i
is a
Moreover, for any f i n i t e subset Pi of
can be chosen
pi
Let
P1
Jli
Xo,mo
p2’
is HA.
i = 1’2,
a h i t muZtipZications
- y2 H-map.
pl, X2’
XlY
Suppose xo,mo
(f.c.n).
be P-Pi
x0
Let
Proposition:
1.
Pi and any
that:
~ ~ ( ‘ i 7 ~f o,r ~ p~ E)
By 1 . 5 . 1 ( a ) t h e existence of
mi
P;
m < n0
so that
$i
is a
pi
-
m
i
H-map i s equivalent t o
Suffices t o f i n d
and W: Xo
i
N
HD($iypiymo) = W(Jli
A
$i)
A
f o r then
X
0
+
Ti
Xo
and
i = 1’2
so t h a t f o r
y2= WA
+ no w i l l be t h e
desired multiplications. Let Xi
and l e t
fibering).
i
8,:
Xi + H t n ( X i. ) i
(n+l)
Yn+l: xi
i = 0,lY2
+Xi
be t h e Postnikov approximations of
be the f i b e r of
8i(Yi+1
-
t h e n-connective
144
Mod p
theory of H-spaces
N
Suppose inductively t h a t multiplications
i = 1,2, were constructed
P-Pi
SO
*
'i ,n'
ai( n ) :
Xo( n )
+
xin)
xi
x
+
xi,
that
Xi
equivalence so t h e r e e x i s t s an i n t e g e r
map
xi
so that
$ln)@in) = Ai
prime t o
1 (n).
and a
P-Pi
P1 tl P2 = k?
As
xO
X1
and
(ai
a r e r e l a t i v e l y prime, and t h e r e exist i n t e g e r s
X2
prime t o
- alAl Let
?,a2
P-Pi) so t h a t
+ a2A2 =
1.
jiYn = ji = Yiain)ai '
( + with respect t o
i ,n
).
N
*
As
Put
",n+l
= +i($i
N
A
$,)A
+
in the proof of 1.5.1 ( b ) one has
The first term can be rewritten as
where
a: X
0
A
X
0
+
(Xo
A
Xo) A
(XoiX
X,)
s a t i s f i e s ah = ( A
-
A
l)Ax 0
xo.
Mod P1 H-spaces
then
S u b s t i t u t e ji =
Yi@:n)(ai
*
un)
.
The first term of
W
i ,n+l
'i,n
decomposes as
[$ Yi@(n) i n i
A
11
xO
As
xt) A xo
is n-connected
[x0(n) A
H t (X,)]
Xo,
n
= 0
and
Bn0
annihilates t h e f i r s t term:
It follows t h a t f o r some N N
HD(JI1,
ulYN9
x?)
Wl,n+l =
Y,
- Wa+l "N
= 0,
-- Yn+lun+l. 0 As WlYN =
w2,N
N
no) = HD($2, I J ~ no). , ~put ~
N
Xo
satisfies
and N
ui = i' ,N'
w
= Wi,"
f.c.r
Mod p
146
primes i n
Pi,
- ji
Ai
for
4.5.3.3. then
and
1
Fix $ : X
= %
Replace
0.
Let
FJ
i s properly chosen
P2
4.5.2
i = 1,2
pi,
and m
- P1
dim < no
be
if necessary s o one m a y assume
mod Pi
*
i'
x
-+
X x Xo 0 (JI-
x
=
-+
Xo
Xo
so that for
for
X.
there exist
i = 1.2
the
r(r-)$) x
'i
are
H-structures
By 4.5.2.1
k
X
admits a m u l t i p l i c a t i o n
X(Pi,+-) a
x(P,,$-)
P2
by
x x x
:
then i n
i s both a mod Pl and a mod P2 H-space
If X
Xo = K(QH*(X,Z)/torsion).
(Al, A2,...)
,pi
i n the inductive s t e p i s then replaced
i ,N
following i s c o m u t a t i v e
'I JI,
any multiple of
i s a mod P1 U P2 H-space.
X
P fI P2 =
By
^x i i s
H * T ; ~ ~=zH~* )( u ~ , z ~ ) .
Corollary:
Proof: -X =
pi
xi
and i f
p E P;
where
a high enough power of t h e product of primes i n
e.g.
Pi f i n i t e .
Pi c Pi¶ by
Ai*ii
can be replaced by
Xi
Npw,
theory of H-spaces
'(Pi)¶
mi
'(Pi)
H-maps.
so that
By 4.5.3.2
one obtains
The genus of an H-space
N
pi
is a
i s an H-space.
X(Pl U P2, $-)
t h e i r pullback
- m132
x
147
H-maps and consequently X
5l
P2 X(P, U P2,
$ I ,
x
mod P1 U P2 H-space. Proof of 4.5.3:
X
Suppose
Then
X
i s an Ho
space.
then
J,
i s a P1
quivalence,
H-space as w e l l as
mod P
Let
is a $: X
By a simple induction 4.5.3.3
Hence
...,k
i = 2,
implies t h a t
H-space f o r every
p.
s Xo = K(QH*(X,Z)/torsion)
P-P1-finite.
= {pi)
i
mod p w o
X
X
where
is a
mod P1
{p2,...,pk} = P-P
i s a mod P
H-space
-
1'
hence an H-space.
4.6.
The genus of an H-space.
4.6.1.
Let
X
be a universal space ( i . e :
Suppose
X
satisfies (f.c.).
Definition:
every set
P1).
P1
!The genus of
universal f o r
X
is the
set G ( X ) = {[Y] A property
[XI E P
4.6.2.
I
Y -p X
for every
p).
P of homotopy c l a s s e s of spaces i s s a i d t o be generic i f
implies
Example:
G ( X ) c P.
By 4.5.3 being an H-space i s generic ( i n t h e f . c .
Combining t h e r e s u l t s of [Zabrodsky]
9
case).
and [Wilkerson] one g e t s a
characterization of t h e genus of an H-space s a t i s f y i n g ( f .c. ) :
4.6.3.
Proposition:
x
xo - Y
X
be an H-space s a t i s f y i n g
if and only if t h e r e ezh?ts an H-space
[Y] E G ( X )
x
Let
x
x0'
Xo
(f.c.n).
so that
Then
148
Mod p
Let
be an H
X
JI: X
0-equivalence
theory of €I-spaces
space s a t i s f y i n g
0
Xo = K(QH*(X,Z)/torsion)
+
(For a f i n i t e abelian group A s a t i s o i n g mA = 0 ) . QH i(X,Z)/torsion
#
L e t .9
Fix an a r b i t r a r y
and l e t
t
be an i n t e g e r
exp r i ( f i b e r
N ( X ) and by
i s t h e smallest i n t e g e r
exp A
be t h e number of i n t e g e r s
ni
If: x
+
XI
f
The main result of [Zabrodsky]
9
is a
p
m
so that
equivalence f o r every
pit}.
states:
4.6.4. Proposition: Let X be an H0 space satisfying
(f.c.).
Then
G(X) admits an abelian group structure w i t h an ezact sequence
n
n
..., l d e t QH ' ( f , Z ) / t o r s i o n l ) .
where a ( f ) = ( l d e t QH l ( f , Z ) / t o r s i o n l , (Note t h a t although
[X,X],
For t h e ( f . c . n )
case,
4.6.5.
Let
dl,...,da I
dl"
n
I
dl'.
..¶d' I &H
i
**
i s not a group
X
im a
is).
6 i s given as follows:
be i n t e g e r s ,
( d i , t ) = 1. Let
: &H*(X,Z)/torsion + &H*(X,Z)/torsion, ¶
(X,Z)/torsion = di
1. Then
ry
class of
$1.
Finally, l e t
0.
=
[x,x],
m
$n(X,Z)
d i v i s i b l e by all t o r s i o n of
n
(f.c.).
given by a pullback diagram
E(d,,
d2,...,dQ)
i s the
149
The genus of an H-space
N
X
4.6.6.
If
X
satisfies
t
t h a t i n t h i s case primes of by
Ilkdim
4.6.7.
for
N
ll
izdim X
exp n i ( f i b e r $ )
and
?
i s replaced
,(PI. !l%e genus of
Example:
SU(n): L e t
( I t i s obvious t h a t i f
(f.c.s).
the same procedure can apply except
i s chosen t o be an i n t e g e r d i v i s i b l e by t h e t o r s i o n and by
H*(X,Z)
(f.c,H)
X
X
be an Ho space s a t i s f y i n g
s a t i s f i e s (f.c.H)
G(X)
-
C(HtN(X))
s u f f i c i e n t l y large s o our observations cover the (f.c.H) case as
w e l l 1. One can choose an i n t e g e r simultaneously f o r function
2 1)'
(Z:/
[X,X], .+
(Zf/
-+
X
t
and H t n ( X )
[Htn(X), H t n ( X ) l t
2 l)',
('ln
t h a t w i l l s a t i s f y t h e 4.6.4 hypothesis f o r all n.
As one has an obvious
and an epimorphism
the number of i n t e g e r s
n < n i -
with
n
&H i (X,Z)/torsion # 0 )
s o t h a t the 4.6.4 exact sequences can be compared
Mod p
150
Hence,
G(X)
X = Sun = H t epimorphism.
2n
Sun
J I : Sun
If
Pt
[SU,,
equivalence,
Pt = { p E PI p l t ) .
-+
If we choose
sum],
n
K ( Z , 2m-1)
- the
t
-+
[SU,,
k = m-1:
Indeed, s u f f i c e s t o show t h i s f o r be a
m,
n
one has even more:
(SU)
-+
f:
i s an epimorphism.
G(Htn(X))
-+
theory of H-spaces
SUklt
i s an
Let
4.6.4 i n t e g e r f o r
By a theorem of Bott ( [ B o t t I 2 )
realizes
Q,H*(SUn, Z ) / t o r s i o n ,
m
r 2m ( f i b e r J I ) = Zm,,
m < n,
Furthermore, t o r s i o n k
H
i n v a r i a n t of
Sum -+
m u l t i p l i c a t i o n by
X
t = II m!,
and
Pt =
mcn
2m
Pt
being a
f
H
of torsion
2m
(SU m-1'
P < nl.
and i t s generator i s t h e
Z ) = Z(P1)!
(SUPl,
{PI
equivalence y i e l d s a Z)
(A,t)
= 1. f
induces a
diagram
f
"m-1
+
"m-1
lk a d
?
E [sum, SUmlt
Now 4.6.7.1 f o r h: G(SUm) + G(SUm-l) onto
image f E
with
X = SU
m'
b u t as
ker h = c ( k e r r),
I
1t -
n = 2m-3
Ht--,:
y i e l d s an epimorphism
[Sum, SUmIt
-+
k e r r = {l, l,.. .,l, dl d E Z/:
SUWllt
2 11.
is
We s h a l l
The genus of an H-space
(1, l,...,d)
show now t h a t Indeed, i f
d
?
fl
lw
where
u
E i m a i f and only i f
((m-l)!)
pl-1
151
d f
fl
((m-l)!).
one has
sh-l
d
c l a s s i f i e s the principal
-
BSU(m-1)
U
bundle
SU(m-1)
N
As
order u =
(IE-~)!
.
aTa = (l,.. ,1, d ) . Put
fo = Ht2m-3(f):
du = u.
Put
fd = Ht2m_l(fd)
Conversely, suppose
-
Sum
f : Sum -+ Sum, a f = (1, l , . . . , d ) .
Then,
-+
homotopy equivalence and by [Zabrodsky] 10
2:
SUmw SUm with
t h e f i b e r of k < 2m-1,
Hth-l(?)
Sum -+ Sum-,,
= fo.
X: SU m
Hence, -+
.. ,1
a f o = 1,. Corollary
1.14:
?-f = jX,
K ( Z , 2m-1).
hence, by [ Z a b r o d ~ k y ]Corollary ~~ 1.8
d i v i s i b l e by
= Sum -+ Sum and
and
fo
is a
There e x i s t s
j: K(Z,2m-l)
-+
ak(jX) = 0
for
degree
(jX) i s
IT
2m-1
sum
Mod p theory of H-spaces
152
It follows t h a t
ker[G(SUm) -+ G(SUm-l)]
FJ
21 and one
Zrm-l)!/
obtains a decomposition sequence
Z2"l fl I
. Consequently,
Cp
G(SU3)
-+
G(SU2) = 0
n
I G(SUn)I
=
II order ( Z rm-1) I 1
m=2 where
-+
t1
i s the Euler function.
G(SU6) = 4.
Finally, t h e epimorphism
(p(120)/2 = 4.16 = 64 e t c .
G(SUb)) -+ G(SUn)
y i e l d s t h e estimate
We dare t o conjecture t h a t t h i s inequality i s indeed an equality.
4.7.
MixinR homotopy types
4.7.1. $i: Xi
Let
Definition: -+
Xo
The r r k h g of
and l e t Xi,
Pi,
pullback o f ' t h e maps of 4.3
1.
Pi (Ji
(J;:
Xi
i = l,...,k
.. ,k
i = 1,.
be a p a r t i t i o n of t h e s e t of primes.
i s t h e space Xi(Pi,
(Ji)
+
be spaces with 0-equivalences
Mix(Xi, Pi, $ i ) obtained as t h e
Xo
(Xi(Pi,$i)
i n t h e notations
153
Mixing homotopy types
4.7.2.
Let
Proposition:
of the same type.
Mix
4.7.2.1.
be universal spaces s a t i s f y i n g
Xi
( f .c .)
Then
(Xi,
Pi, Jli)
-
xi
Xi
i s a mod P
'i
I f for every
4.7.2.2.
Mix(Xi, Pi,
ai)
i
is an H-space.
If there e c i s t s a space
4.7.2.3.
H-space then
i
so t h a t f o r every i
X
Xi E G ( X )
then Mix ( X i , pi, Jli) E G ( X ) .
Proof:
4.7.2.1 P-Pi
N
If
follows from t h e following observations:
f i : Xi
Xo
then t h e pullback of t h e induction f o r
i = 1,...,k,
is
fils
Pi
{Pi)
d i s j o i n t s e t s of primes
Xi.
equivalent t o
Indeed, by
k = 2
J f2
P1 c P N
P-P1
u
replace
-
implies
P2
P2.
For
X1,X2,
as above),
'xo
P-P2
x2
x12 NP1
X
2'
similarly
k > 2,
given
X1,.
. .,%,
fl,f2,
P1,P2
by
X12
f12 = f2gl,
B1 U P2
X12-P2 fly
- .. , f k ,
X
1'
ply
( t h e pullback of
f2gl
is a
- .. ,Bk fl
and apply t h e induction f o r
and
f2
1 54
Mod p
.,$, PIU P2,
X12,
X3,..
f12,
f3,...,fk
theory of H-spaces
. .,Pka
P3,.
i s t h e pullback of
f3,.
f12,
.., f k :
f2a...sfk
fla
X
The pullback
and
x
xi,
i
of 2,
pi
X
wP,
rh, Xi
B2 X12
4.7.2.2
i = 1,2.
X
follows from 4.5.3 as
Xi
3
X
imply
a mod Pi
H-space,
i hence 4.7.2.4
4.7.2.3 i s immediate from 4.7.2.1,
f o r all p.
mod p
follows from 4.3.3.
The following r e l a t e s t h e concepts of mixing and genus of an
Ho
space : 4.7.3,
Let
Proposition:
JI: X
arbitrary 0-equivalence integer
if
i0 80 t h a t
be an Ho
X
-+
X
0
= K(n,(X)/torsion).
Fix an
There e x i s t s an
i s a P1 equivalence f i n
= (X0l)$
$,
space s a t i s f y i n g f f . c . ) .
0
dim ( N ( X )
i n the
case) then
(f.c.H)
e x i s t s a P2 = P-P1 eqtrivalence
4: Xo
[Y] E G(X) -+
Xo
i f and only i f t h e m
of the form
4 = r(IJ A
(in
the notations of 4.41 so t h a t Y w
Proof:
MiX(X,
Obviously
P1’ P2; 4JIX0
JIXo
Mix(X, X; P 1
4.
$,
0-equivalences
x;
Conversely, suppose
[Y] E G ( X ) .
As t h e ( f . c . n ) case i s simpler we
s h a l l prove t h e (f.c.H) case.
As i n 4 . 1 . b f a c t o r s through spheres.
t h e r e e x i s t s an i n t e g e r H t N ( X ) ( S ) , where
= X $ 0
S
ho
so that
X 01: Xo
-b
xo
i s a product of odd dimensional
thus y i e l d s a commutative diagram
Mixing homotopy types
155
S +H t N ( S ) _ _ j X
+O
I
By 4.6.5 and
4.6.6 there e x i s t s 4 = I'(I-) so t h a t
homology approximation i n a
J,
-
Y
x
4 p u l l back diagram.
xO
see t h a t t h i s i s equivalent t o t h e following:
4.7.3.0 -2
1 X-
N
pullback
1'
pullback
. JIY
Consider now:
JI:
Y
f
p2
Y
X
h
i s given as a One can e a s i l y
Mod p
156
theory of H-spaces
By the uniqueness of the 4.3.1 decomposition one obtains:
(OJ, 4.7.3.1
yP2
Y
,
Y U
4.7.3.0 y i e l d s N
Y
h W P
S1
a'
I
a''
i = i,i, X
h
FJP
s-
at
1
Hence : 4.7.3.2
4.7.3.3
J
xO
157
The non c l a s s i c a l H-spaces
.
( @ $ )I' can be replaced by p1
(4.7.3.2).
-4.7.4.
( @ $ )If p2
By 4.7.3.1 one can replace
Corollary:
Hence,
;) 0 2
by
.
By 4.7.3.3
which i n t u r n can be replaced by
$u"
Y = Mix(X, X; Ply P2;
To be (of the homotopy type) of a f i n i t e CW loop
space i s generic, hence the mixing of f i n i t e d CW loop spaces i s a f i n i t e CW loop space.
Proof: Y
FJ
If
(a s a t i s f i e s
[Y] E G(0X)
Mix(X,X; a$, $; P1,P2).
One can a l w a y s choose
= r(I-)
t o be a loop map.
and
A
procedure t o construct
4.8.
-
By 4.7.3
(f.c.H)).
STX
$:
- 0
OXxo
X 1 are obviously loop maps and t h e 0
Mix(nx, QX; Ply P2; $JlX0.
qh0) i s deloopable.
The non-classical H-spaces and other applications Until 1968 t h e only known f i n i t e CW H-spaces were t h e c l a s s i c a l ones:
Those of t h e homotopy type of L i e groups,
S7
and t h e i r products.
I n 1968 Hilton and Roitberg have discovered t h a t bundle over
S7
c l a s s i f i e d by
H-space and a non c l a s s i c a l one. ( [ S t a ~ h e f f ] t~h)a t
E
7w
7w E
3
n6(S. )
-
E
7w
-
the
S
3
Z12 (EW = S p ( 2 ) ) i s a n
It was l a t e r shown by Stasheff
i s of t h e homotopy type of a f i n i t e dimensional
loop space.
observation follows from 4.7.4. discovering an (e.g:
By 4.6.7
-
The l a t t e r i s obviously a source f o r
family of non c l a s s i c a l f i n i t e dimensional loop spaces
G(SU(n)) i n i t s e l f i s a very r i c h source for such
Mod p
158
theory of H-spaces
I n f a c t till now we know of no f i n i t e dimensional loop space
spaces).
The same holds f o r
which i s not i n t h e genus of a L i e group.
mod 2
H-spaces where we know of no non-classical f i n i t e dimensional ones. s i t u a t i o n i s d i f f e r e n t as far as general H-spaces a r e concerned. Sa+'
(see [Adams] ) was t h e first t o observe t h a t 2
The
Adam
i s a mod odd H-space.
(Actually, from t h e point of view of 4.5.2 t h i s follows from t h e much e a r l i e r observation t h a t and t h e r e e x i s t s a map mod odd H-structure).
2[1 , I ] = 0 , S2n+l
s2n+l
E
I
IT^^+^
s2n+l
( S2n+l)
t h e generator, which i s a
of type 2,l
This i n t u r n yields a huge family of non c l a s s i c a l
H-spaces :
4.8.1.
nl, n2,...,n
, Q)
H*(%
1
... 5 nr
r y n1 - n2 -<
....,xn
= h(xn 1
Z,Q)
x
1
Mix($
-
5
x
5,
m e r e always
1
i
2m +1
N
S = VS
so that
1
H*(G,Q),
-
x
i*e:
'(5 , Q).
xn
r
nJ-oddy
n. E H
e x i s t s a f i n i t e product of odd spheres
H*(%
a mod P H-space of type 1
be a s e t of odd primes,
Bel
Let
-0
a Lie group.
G
Xo = K(n,(G)/torsion)
Let
a2:
G
-0
Xo.
Then
G; P1,P-P1; $1,$2) i s a non c l a s s i c a l H-space.
1
4.8.1 i l l u s t r a t e s how t h e construction of mod p H-spaces leads t o t h e construction of non c l a s s i c a l H-spaces. method t o construct
mod p
We s h a l l present here a
H-spaces which i s a general form of a
construction described i n [ S t a ~ h e f f ] ~ . I n t h e following proposition by a f i n i t e s t a b l e complex w e mean a spectrum
k
= {I%)
connected we w r i t e
where
K
i s a f i n i t e complex.
H ~ ( ; O= H ~ + ~ ( K ) . dim ic
= dim
If
K-S.
K
is
s-1
159
The non c l a s s i c a l H-spaces
r = EBm
D = C mBm, m
Put 1 < 11
be a f i n i t e stable complex, Hodd(?,Z) = 0
Let
Proposition:
4.8.2.
Bm are the B e t t i n d e r s of 2.
where
If
is an odd integer satisfying 2(D +
r e $ )
then there e d s t s an R - 1 a: SX +
k
< pI1
+
p-2
connected f i n i t e
H-space
mod p
X
o f degree zero o f the stable complex SX = IZnX)
with a map
so that the
composition
is an isomorphism.
Proof:
K, N-large.
complex chosen map
Represent
n
k
x Hkk(fr) -+ HI1,,(i)
HE (?) k
-
-
for s u i t a b l y
for some
P
induced by
dimensional
j , = QnZK
The i d e a i s t o show t h a t
w i l l have t h e property
Hak(?)
4N
2N-1 connected
by a
Add:
i
x
3
j,
k.
The
can be
mod p
n
deformed t o a mod p
H-structure
Hllk(?)
The l a t t e r s t a g e i s q u i t e simple, as space
Y,
given
hf;l i s a mod p Ha,(%)
ej HI1
P
2k
h:
Him(
x
a)
Hak(i)
i s p-universal f o r any then i n
H-structure for X = H l l k ( X ) .
(2)
Hkk(i[) = X.
H l l k ( k ) + H!Lk(X),
(depending on t h e choices of
The observation
n
and
k)
is a
160
theory of H-spaces
Mod p
H. (ft,Z
consequence of t h e a l g e b r a i c observation
To see t h i s one has t o analyse Y
Let [Browder]
4'
HS(Y,Z
P
) = 0
-
( a ) o*: Q H ~ ( Y , z ~ )
w
P
) = 0
for
k < i
Zp)
H"(QnCK,
be an H-space with theorem 5.14
1
PH'-'(QY,Z
By
s < 2(m-1)p+1.
for
P
s < 2m-1.
for
A simple Hopf algebra argument y i e l d s
Again by [ B r ~ w d e r ] ~Theorem , 5.14 (c)
w
&H~-'(QY,Z~)
Finally
(a)
-
s-2 QH
w
P H ~ - ~ ( Q ~ zp) Y,
Putting (a)
P H ~ - ~ ( Q ~zY , P
2
( a Y, zP
for s < 2(m-l)p
for
( d ) t o g e t h e r one has
Q H ~ ( Y , Z ~ fil ) H'(K,Z~)
H'(Y,Z H2'+'(Y,Z
P
) = 0
P
for
) = 0.
for
s < 2N,
s <
4N,
s < 2(m-l)p+2
5 2k.
The non c l a s s i c a l H-spaces
and
s < 4N
for
s-2n
(or equivalently
s
-
2n-1 < 4N
-
2n-1).
161
Substitute
2s
for
one o b t a i n s 2s+2n 2s+2n-2N 2s-1 2 n + 4 , Zp) = H (K,Z ) = H (K,Zp) QH (Q P
for
2s-1 < 2(N-n)p-2
and
2s-1 <
4N
-
As
2n-1.
QH2s+1(k,Z
P
) = 0
one o b t a i n s
Q,H2s(Q2n+4, Z ) = 0 P Choose
then
for
4N
-
N
and
n
s < (N-n)p-1
if
and
so that:
2n-1 = 2N + il > (k+l)p-2 = 2(N-n)p-2,
hence,
2s-1 < ( ~ + 1 ) p - 2 . w ~ ~ ( Q z~ += o~ f, o r P Let
A
s < 2N-n
2s < ( ~ + 1 ) p - 2 .
be t h e e x t e r i o r a l g e b r a generated by
QH2s-1(8n+hy Zp)
2s-1 < (E+I.)P-~. Then
2k = 2 dim A < p .(11+1)
By t h e hypothesis A
N
H*(i22n+%,
Zp)
through
X = Hllk(02n+4) -p
1
D + llr < $p(E+l)
dim p(ll+l)
St?"+lY).
-
2)
-
-
2
2
and consequently
and
One can e a s i l y argue t h a t
implies t h a t
dim
fc
< (ll+l)p-2-1
and hence
Mod p
162
4.8.3.
theory of H-spaces
Let
Realizing Hopf algebras:
Suppose
i s odd t o r s i o n free.
H*(G,Z)
P1 and Hopf aZgebra s t r u c t u r e s
6,
there exists an H-space p E B1
H*(6,Z
Proof: Let
give
s o that
p E Ply
= H*(G,Z)
H*(G,Z)
and for every
N
i s 0-equivalent t o a product of odd spheres
G
k = .IISi 1=1
p '
( TI k j ) j
A
x2m+l ( II
mod
P1 equivalent t o S
a multiplication
so t h a t
H*(g,Z
P
S =
k 2mi+l ll S i=1
.
one can e a s i l y
) = A P
as a Hopf
(This i s being done by r e a l i z i n g t h e reduced coproduct
p E El.
X hi+l+
-+
qP f o r Ap = H*(G,Zp)
as Hopf aZgebras.
be an H-space,
algebra,
q1: i
) = A P
Oiven a f i n i t e s e t of odd primes
a +1
Si
i
P
be a f i n i t e dimensionaZ H-space.
G
- x2m+1 ' l-l ' x h + l
i j )-+ gi).
v i a a map
One has 0-equivalences
j
a multiplication
i
1
and m u l t i p l i c a t i o n s
H-map and
I),a
u-m
JI,:
= Go y
K(n,(G)/torsion)
G -+ Go.
By 4.5.3.2
one can f i n d
*
for LI
S
SO
(for
G)
that m
H*(C (for
Z
1' P Go)
= H*(j,Zp) SO
that
H-map.
Mix($, G; Ply P-P1;
IJJ~,
J12)
i s t h e d e s i r e d H-space.
JI,
p E P1 is a
i,-
m
163
Chapter V
Non s t a b l e BP resolutions
Introduction The Adams-Novikov s p e c t r a l sequence proved t o be e f f e c t i v e i n sane
computations i n t h e s t a b l e homotopy groups of spheres (see [Zahler]).
In
t h i s chapter we attempt t o advocate f o r t h e use of t h e BP-spectrum as studied by Wilson (e.g. [Wilson] and subsequent papers) f o r some non s t a b l e problem. The connection t o H-spaces l i e s with t h e f a c t s t h a t our model spaces
a r e H-spaces and some H-spaces invariants a r e occasionally used and t h a t some of our applications deal with H-spaces.
We give here a very loose account of t h e prospective theory and our applications are just a scratch on the surface of p o s s i b i l i t i e s and w e dare s a y t h a t already here i t s efficiency compare t o the c l a s s i c a l obstruction theory can be observed.
(It does not imply however, t h a t
some ugly arithmetic computations can be always avoided). Throughout t h i s chapter w e use many of Wilson's r e s u l t s without proof. A s we a r e t o s t a y within t h e category of CW-complexes of f i n i t e type
we adopt t h e Brown-Peterson o r i g i n a l construction of the spectrum BP f o r our nonstable model.
(See [Brown Peterson]).
t h e r e s u l t s obtained t h e r e .
W e use f r e e l y some of
164
Non s t a b l e
5.1.
BP
resolutions
K i l l i n g homology p-torsion
5.1.1.
Let
Proposition:
e x i s t s a space fa)
F(X) = F
H,(F(x), Z )
f b ) n,(fiber h)
fcl
be a space.
X
(PI
For a given prime p
and a map h: F(X)
(X)
there
so that:
+ X
i s p-torsion f r e e .
is torsion f r e e .
There e x i s t s a commutative diagmm
A
fd)
h: F(X)
over
-+
x.
X
does not dominate
i.e.
mod p
any non t r i v i a l space
If in the following commutative diagram
\ Fj: // gl
g2
’
y1
y2
X
g2gl
Proof: hn: Fn(X)
-+
is a mod p equivaZence, then s o are
F( X )
X
gi’
i s constructed i n d u c t i v e l y as follows :
is constructed so t h a t
and t o r s i o n f r e e f o r a l l
m,
am(fiber hn)
H~(F,(X), Z)
and one h a s a commutative diagram:
is
i s p-torsion
0
Suppose for
m 2n
free for
m 5n
165
Killing homology p-torsion
Let
En:Fn(X)
-+
K(Hn+l(Fn(X),
Z), n+l) yield an isomorphism on
N
homology in dim n+l. Put An+l
= Hn+l(Fn(X)y Z ) / # p torsion,
N
kn : Fn(X)
-+
K(An+ly n+l) induced by
cn
and the quotient.
Let An+1
N
be a free abelian group, an+l: An+1 -+An+1 a surjection and an+l 8
p an
isomorphism. Let F (X) complete the following into a pullback diagram: n+l
Using the Serre exact sequence one can see (mod # p that Hn+l(fiber kny Z ) = 0 and Hn+l(~n+ly Z ) Hm(hn+l,n, Z )
is an isomorphism for m 5 n
torsion)
is an isomorphism. As
Hm( Fn+l(x),Z )
is p-torsion
free for m 5 n+l. -
put hn+l
- hnhn+l,n and one has a fibration Fiber hn+l
,n
-+ Fiber hn+l
-t
Fiber hn-
Fiber h = Fiber K(an+l, n+l) = K(ker a n+l). n+l’ n+l ,n
ker a cA is n+l n+l
166
Non s t a b l e BP r e s o l u t i o n s
free, hence,
m 5 n,
TI
n+l
n m ( f i b e r hn+l)
can be constructed a l t e r n a t i v e l y by composing
’H,+~.)
N
(Yn induced by any s p l i t t i n g An+l
Fn+l
+ pHn+l((Fn(X), Z ) ,
so that
‘n+l
(X)
n+l
B Z
p
Let
+
n + l ) be any epimorphism of a free group Fn+l (X) i s then t h e pullback
i s an isomorphism.
i s t h e f i b e r of
Ynkn
- K(Gn+l.
n+l): Fn(X)
x
K(in+l, n+l)
n
hence,
Fn+l =
X
hn)
( f i b e r hn+l) = nn+l(fiber hn+l,n).
Fn+l (X)
“n+l: An+l
n m ( f i b e r hn+l) =
i s free:
F,+,(X)
-jp-p
(XIE w P- . F(X)
x K(n,(fiber
Fn(X)
x
hn+l))
x
K(pHn+l(Fn(X),Z),n+l)
n
K(An+l, n+l).
K ( A ~ ,n+l)+ ~
+
Now
X
x
An+l
= n n+l ( f i b e r hn+l ,n
K(r,(fiber hn) x K(in+l,n+l)=
The following commutes:
Killing homology p-torsion
167
X satisfy (a) - (c). F(X) , h = 1Jm Fn(X) , l$.m h n
To see
(a), suppose
one is given
g2
F(x)
\ I. X
2 y /
a p-equivalence. Decompose h, hi into a Postnikov decomposition:
Hn(hi ,n,n-l'
Z) is surjective (fiber hi,n,n-l being n-1 connected) l
and if one supposes inductively that
gi ,n-1 axe p-equivalences
Non s t a b l e BP
168
Hn(gl,n’
i s a mod p
Z)
isomorphism.
is
mod p
exact it f o l l a i s t h a t
is
mod p
surjective.
Z)
n
l,n,n-1
it follows t h a t fiber hl ,n ,n-1
5.1.1.1.
-+
,Z)
n,n-1
_.
n,n-1
mod p
type of
z). = K(T ( f i b e r h) , n) n
h: F(X) + X
t o r s i o n free b u t not both, otherwise t h e
mod p
mod p
or
is
H (F(X), Z )
n
minimality ( p r o p e r t y ( d ) )
i n v a r i a n t version of property ( c ) i s t h e weaker
\/ X
The c o n s t r u c t i o n of construction
F(X)
h: F(X)
-+
X
i s not unique.
n
:
H ~ ( F ~ (,xz))
However, once a 5.1.1
i s chosen one o b t a i n s i n v a r i a n t s N
~1
,Z)
a r e p-equivalences.
sn(fiber h)
The
n ,n-1
equivalence
concerned we can f i n d models with e i t h e r
w i l l be l o s t .
Hn(fiber h
Z)
fiber h
gl,nS g2,n
and
As f a r as t h e
-
+
l e f t inverse
Hn(fiber h2,n,n-ly
y i e l d s a mod p
fiber h
Remark:
mod p
= K(n;l(fiber h l ) , n),
g l,n
-+
Hn(fiber hl,n,n-l
Z ) “p, Hn(fiber hn,n-ly
H ( f i b e r hl,n,n-l,
fiber h
As
Again, having a
H ( f i b e r hn,n-l, n
As
resolutions
/# P t o r s i o n =
An
-f
An.
K i l l i n g homology p-torsion
5.1.2.
169
Proposition:
( a ) Let
H*(Y,Z) be p-torsion f r e e .
l i f t s t o a map
f: Y
+.
f: Y
-t
X
Moreover, if f o r
F(X).
Hn(Y,
an#: H n ( Y , A,)
Then m y map
xn) (which is onto)
one chooses a l e f t inverse (as a f m c t i o n of s e t s ) an"Jn= 1 then
rH ~ ( Y An), ,
Jn : 8 ( Y , x n )
1 can 1 3
be chosen i n a unique way so t h a t [u ] = J [k n n n n-1 n-1
(p:Y
Fm(X)).
+.
*j; then any map
(bl If g: X gh: F(X)
-f
X
* ?) is a
(c) If X is a mod p mod p
Proof: to
Fn(X).
p
g:
F(X) + F(?)
(Zifting
equivalence.
H-space then s o is
F(X)
and h
is a
H-map.
(a) Suppose
i s given.
J
Suppose i n d u c t i v e l y
f
lifts
One has t h e following commutative diagram
K(an+l, n+l) Fn(X)
kn
b
~.( ; ni+~ l) +~,
, which y i e l d s a map and
K(an+l, n + l ) ) :
-
?n+l. Y
-+
F
n+l( X ) ,
(Fn+l(X)
- the
pullback o f
k
n
Non s t a b l e BP
170
Now any l i f t i n g J:
fA+l: Y
-+
F
n+l
[ u ~ + ~ ? A + ~=]Jn+l[kn3n 1
(X)
resolutions
?n
of
w i l l satisfy
induced by
[ u ~ + ~ ? =~ Jn+l[kn?n] +~]
=
[“n+l?n + l 1. fn+l =
Hence, i f =
(Pn
x
A.Y
(p
N
x Y x U ~ + ~ ^ ~ ~ + ~ ) A : Y: ,
A: Y
x
m u l t i p l i c a t i o n o f functions.
AS
-+
H”+’(Y,z)
N
A: Y -+ ~K(A,+~, n+l) l i f t s t o
A: Y
r e p r e s e n t s a self homotopy o f
un+lfn+l
core1 h
n+l ,n
Of
fn+l
nK(Tn+ly
-+
Y
+
n+l),
PK(An+l, A-Y
i s p-torsion
O K ( A ~ + ~n,+ l )
n + l ) then
- t h e pointwise f r e e any map
and
which induces a homotopy
and g
(b)
Suppose i n d u c t i v e l y
pullbacks :
Fn(X)
Fn (XI).
Given t h e map o f
Killing homology p-torsion
hn+l ,n
h;+l ,n
\ As
an+l Q Z P
p-isomorphism. and so is
8 Zp
and
are isomorphisms Hn+l
gn+l, Z )
( A
is a
Consequently K(Hn+l(bn+l, Z), n+l) is a p equivalence
gn+l*
(c) Let
p:
X
x
X
+
X be a mod p H-structure. By ( a ) one has
BP
Non s t a b l e
172
uii,
As
i s a p-equivalence, hence, by ( b ) F(p)u i s a
p-equivalence, mod p
resolutions
H-structure f o r
mod p
is a
F(p)ui F(X)
and
is a
h
H-map.
5.1.3.
Proposition:
of 5.1.1 i . e . :
h: F(X)
~f fi: Y
-+
-+
X
x satisfy
F(X), h
then there e k i s t s a mod p
and h g -
^h.
Proof:
(a)
-
( d ) with
equivaZence
Y,
g: Y
h replacing
A
5.1.1 ( a ) and 5.1.2 ( a ) imply t h e e x i s t e n c e of
Now, 5 . 1 . 1 ( b ) and ( c ) f o r
Y
imply t h a t t h e
a r e all i n t e g r a l of order power of
-
i s characterized by properties ( a )
p
k
hg
i n v a r i a n t s of
and t h e f a c t t h a t
i s p-torsion f r e e implies t h e e x i s t e n c e of
g,
v : F(X)
-+
Y
F(X)
Y
H,(F(X), $v
-
-
h.
X
-+
Z)
h.
Considering
A
by 5.1.2 ( b ) both v
5.2.
and
i s a p-equivalence and by p r o p e r t y 5.1.1 (d) for
gv g
Wilson's
a r e p-equivalences.
B( n , p ) ' s Let
B(n,p)
i s t h e space B(n,p) = F ( p ) ( K ( Z , n ) ) = F ( p ) ( K ( Z p , n ) ) .
mod p
be an i n t e g e r and
p
a prime.
Definition:
Yn
n
-
5.2.1.
corresponds t o
Y
Wilson's (B(n,p)
i n [Wilson]) 5.1.1 and 5.1.3 y i e l d t h e following
c h a r a c t e r i z a t i o n of
B(n,p):
(d)
Wilson's
2.2.2.
The mod p
Proposition:
( a ) B(n,p)
(bl
type of
an(B(n,p)) 0 Zp
is characterized by:
i.e.:
am(B(n,p))8 Zp = 0
0
is p-torsion f r e e .
H*(B(n,p), Z )
is p - t o r s i o n f r e e .
( c ) a,(B(n,p))
(a)
B(n,p)
is an Ho
(el
B(n,p)
does not dominate
is a mod p
g2gl
#
173
B(n,p)
connected mod p,
is n-1
for m < n,
B(n,p)'s
space. any non t r i v i a l . space, i.e.:
mod p
equivalence and &(Yl,
Z )
P
# 0 then
gi
are mod p equivazences.
Proof: d e f i n i t i o n of
We f i r s t show t h a t
F
(P 1
B(n,p)
satisfies ( a )
(5.1.1) ( a ) holds f o r B(n,p).
s a t i s f i e s 5.2.2 ( b ) , by 5.1.1 ( b ) and ( c ) and ( d ) .
Y1As
g1
By 5.1.2 ( c ) B(n,p)-
g2
B(n,p)
Y2
g2gl
an(B(n,p))/(# p)torsion = Z
n n ( g ) / ( # p ) t o r s i o n = Z. Then t h e r e exists
B(n,p)
i s a mod p
- mod p either
-
(e).
By t h e
By 5.1.1 ( a ) ,
B(n,p)
satisfies 5.2.2 ( c )
H-space so i f
M~
Y1
an(Yl)/(# p ) t o r s i o n = Z
or
equivalence
B(n,p)
x
With no l o s s of g e n e r a l i t y assume t h e l a t t e r .
B.
Non s t a b l e BP
174
By 5.1.1
(a)
are
i2,p2
equivalences and z,,(Yl,
p
i: Y
p-torsion f r e e and l e t isomorphism of a2
a1 = 1.
n
n
.
the
i n v a r i a n t s of
they a r e of order
a mod p
Now
k
mod p
a
n ( ~ ) / # pt o r s i o n n
9
)Z
a r e i n t e g r a l of f i n i t e order and we m a y assume
pry
as
H,(B(n,p),
i s p-torsion f r e e
Z)
equivalence and by property ( e ) f o r
(al)*h
Proposition:
Let
Then
X
be an Ho
We first show t h a t
u
is a
vu i s
mod p
space with H,(X,Z)
and a,(X)
X w IIB(niy p ) . P i
F i r s t note t h a t being an
i n v a r i a n t s of
Y
(b)
(ai)*).
Ho
space and having p-torsion f r e e
homotopy groups i s equivalent (up t o mod p the
B(n,p)
is
(a2)*(al)* i s a homotopy equivalence so by 5.1.2
p-torsion free.
Proof:
P
( a ) and 5.2.2 (b). By 5.2.2 ( c ) and ( d )
equivalence (and obviously s o a r e
5.2.3.
w
a
I.
l i f t s by 5.1.2
(a2)&
lifts.
Then n n ( Y )
(e).
+ K ( ~ ~ ( Y ) / ( p# ) t o r s i o n , n)
2
1
One obtains
Indeed k
-
a
Let
P
1
satisfying ( a )
Y
The
2 ) = 0.
I/(# p ) t o r s i o n = Z w i l l imply Y
a l t e r n a t i v e assumption n n ( Y Conversely, given
resolutions
equivalence) t o t h e f a c t t h a t
are i n t e g r a l of order power of
X X
t h a t t h e following commutes:
is a
mod p
H-space.
p.
Suppose inductively
.
Wilson's
175
B(n,p)'s
xvx 0
,
I
/ 0 0
hn ,n-1
0 0
0
I I
/ , , -
x x
x
-
The obstruction f o r the existence of hnF
-
mn i, i s t h e c l a s s
and consequently knmn-l
i s monic
1E
'"n-1
[X
mn' X x X A
f r e e , hence
and mn
exists.
HtnX,
hn,n,l m
K(nnX, n+l)].
X,
has an order power of
too has order power of
,. mn - l - *
-+
p.
But
Now,
and as
p
H*(X
A
lim m * X x X 4 n'
-+
X
kn
H*(A,
X, Z)
= mn-l,
n X) n
i s p-torsion
i s a mod p
H-structure. Suppose inductively
X
2
W
p Then
k
lifts to
Bk"
x W(ni, p )
and
s a t i s f i e s t h e hypothesis of 5.2.3 as well. F
k
is
m-1
connected.
i
(K(amk, m ) ) = B(m,p) (PI rank ( rm2) = B(m,p) -+ K(nmft, m) l i f t s t o
km -+ H t m i
= K(rmk, m)
On t h e o t h e r hand
t o obtain
176
Non s t a b l e
and by 5.1.2
^x
gkn x P
M
(b)
mod p
resolutions
It follows t h a t
equivalence.
rank n m ( i ) f i b e r g2,
i s at l e a s t
f i b e r g2
is
g2gl
BP
X w f i b e r g2 x IIB(ni,p) x B(m,p) i
m
connected.
and
Xw
The p-equivalence
P
llB(ni,p)
follows from a simple limit process. I n [Wilson] it was a c t u a l l y shown t h a t type o f ) --loop
spaces, hence, an
homotopy and homology i s
mod p
Ho
5.2.4.
Examples:
n
mod p
--loop
space.
a r e f r e e algebras
H*(B(n,p), Z ) P
mod 2p-2.
The 5.2.2 c h a r a c t e r i z a t i o n of
f o r every
B(2,p) = K(Z,2)
are (of the
space with p-torsion f r e e
equivalent t o an
It w a s a l s o shown by Wilson t h a t
on generators of dim
B(n,p)
p.
B(3,2) = SU,
B(n,p)
easily yields:
B(4,2) = BSU.
B(3,3) = Spy
B(4,3) = BSp. I n g e n e r a l it w a s proved by Peterson ( s e e [ P e t e r s o n ] ) P-1 t h a t S U M II Xi(p), Xi(p) mod p indecomposable and 2 i connected. P i=l
For 1 1. i 1. p-1 Yi(p)
2i+l
j.3.1.
free.
Lemma:
Y
of maps
-
(M + Y )
BSU
1 1. i
-
P-1 ll Yi(p), P i=l
5 p-1.
, B(n,p) 1.
Let
an Ho
Similarly
B(2i+2, p ) = Yi(p)
connected and
The groups [
5.3.
is
B ( 2 i + l , p ) = Xi(p).
be spaces,
M, Y
space.
Let
(see 4.21.
H,(M,Z)
and
n,(Y)
[MYYIP be the s e t of p-homotopy classes
Then H,(
,Q):
[M,Y]
P
+Hom(H,(M,Q),
1-1.
Proof:
W e replace
fi
p-torsion
by a p-equivalent space
H, ( h ,Z )-tors i o n f r e e : I n d u c t i v e l y c o n s t r u c t
M
with
H,(Y,Q))
l i m Mn = M
..
--ir
4 . .
-
M
T-l - ..
,
hn
n
.A
Mo = M
Ikn
J.
K ( t o r s i o n Hn(Mn-l,
hn
-
t h e f i b r a t i o n induced by
m 5n
for
Y
by a p-equivalent space
JI: Y
JI': Y
2?.
[M,Y]
[MYYIP=
P
-+
implies
-
-
$'fh
To show t h a t
homotopy
Htn(fl)
-
Indeed, a homotopy so t h a t
w
k
Htn-l(f2):
*
Htn(fl)
Htn-l(fl)
-
Z)
is torsion free
similarly, r e p l a c e
torsion free:
$1
= Y({p),
Let
i n t h e terminology
i s i n j e c t i v e and t h e n considering
JI'gh
[h,?]
and
f
which w i l l show
N
P
@;,
i s i n j e c t i v e (and moreover, -+
Hom(H,(h,Q),
Q ) = H,(f2,
H,(fly
+ Htn-l(?)
Htn(f2)
T,(?)
, H(?,Q))
i n j e c t i v e we s h a l l show t h a t i f Htn-l(fl)
Hm(Mn,
M.
is torsion free.
Put
Hom[H,(M,Q), H,(Y,Q)]
[h,?]).
-+
We s h a l l show t h a t
Hom(H,(k,Q)
= H,(g,Q)
%
with
,Q)
H,(
H,(f,Q) that
Y
K(a,(Y)/torsion).
of 4 . 3 ,
[h,Y]
h:
H,(~,z)
and consequently
n)
from t h e p a t h space f i b r a t i o n .
kn
are p-equivalences and so i s
hn
Z),
Q)
Htn(f2).
then any homotopy
can be ( e s s e n t i a l l y ) l i f t e d t o a
and one can pass t o a l i m i t
-
is
H,(?,Q))
Htn,l(f2)
induces a map
fl
w:
h
f
2'
-f
K(nn(i),n)
This homotopy covers t h e homotopy
178
Non stable BP resolutions
Htn-l(fl) = hn,n-l(w (hn,n-l: Htn(^Y)
+
II
-
Htn(fl))
hn,n-1Htn(f2) = Htnml(f2)
Htn-l(i)). A
As
?
and there exists
is an Ho space Y wo K(a,(Y))
n) so that the composition
u: €Itn(?) + K(rn(?),
is a 0-equivalence. One can follow u by an 0-equivalence if necessary so that one may assume yl = h * l ,
If q: K(rn(Y), n) (W Y
Htn(fl) = q(w
H,(Htn(fl),
x
x
0
#
h E
Htn(Y) + Htn(Y) is the principal action one has url = Xpl
f )AA) 1 M
Q) = HY(Htn(f2), Q) H,(dtn(fl),
as H+,(~I,Q), n , ( i )
[k,
Z.
+
up2.
As
are torsion free
K(rn?, n)]
Hom(Hn(h,Z), Hn(K(an?, n ) , Z ) = rn?)
-+
+
are injective and uHtn (fl)
Hom(Hn(?,Q),
Hn(K(r
p,
?
uHtn(f2).
N
corresponds to a homotopy w *
Y
Htn(fl)
The homotopy Htn(f2)
homotopy Htn-1(f2)
-
-+
n ) , Q) = nn? 8 Q)
Hence 0 = hw E H"(i, rn?). As the latter is torsion free w
core1 h n,n-1
But
Q) = H*(dtn(f2), Q).
Htn,l(fl).
-
N
-*
and
Htn(fl)
w%tn(fl)
N
Htn(fl)
covers the
3.3.1.1. On:
Remark:
-
Htn(fl)
H t (f2) t e c h n i c a l l y does not cover
n
-
Qn-l: Htn,l(fl) homotopy
Y
k
-+
L II,(y) < n=O
On-1
+
It covers t h e l a t t e r followed by a constant
Htn,l(f2).
with non zero l e n g t h
-
SO
n
where n-1
II(n)(y) = L
that i f
= $n-l + ka
II : Phn,n-16n
Q,, t n t o satisfy
However, one can e a s i l y choose
R+.
m
The homotopy described i n t h e proof of 5.3.1
On + k
IIrm(y) then
covers
,(n)
n*-
kII(n-l)
As a d i r e c t consequence of 5.1.2 ( a ) and t h e d e f i n i t i o n of
B(n,p)
one has :
5.3.2. B(n,p)
Lemma: +
Let
0:
[X, B(n,p)l
Proposition:
P
.
If
If
H,(X,Z)
is surjective.
-+ r[
i
For any group
a = rank G Q Z
, Zp)).
.kt X be a space with H,(X,Z)
Then there e ~ s t sa: X Proof: -
be induced by
Zp)
$(X,
(yielding an isomorphism o f Hn(
K(Zp, n)
i s p torsion free then u
5.3.3.
+
H,(X,Z)
B(ni, p )
G
F
(P 1
s o that
p-torsion free.
(K(G,n))mp B(n,p)'
i s p-torsion
is surjective.
H*(a, Z ) P
free
where
i: X
-+
K(H,(X),
Z)
a
lifts to
a: X
rank H,(X,Z
P
).
-+
F(p)(KH,(X,Z)) w IIB(n,p) n ,
a
n
Q Z ) = P
i s t h e desired map.
5.3.4.
Proposition:
H,(M,Z)
and n,(Y)
surjection H*(f,Z)
an = r a n k ( H , ( X , Z )
Given spaces
L, M
are p-torsion free.
and an Ho If
f: L
then f o r sane p-equivazent
+
space Y. M
? of
yieZds a
Y
Suppose
180
Non s t a b l e .EP r e s o l u t i o n s
f*: [M,Y] + [L,?]
is s u r j e c t i v e . Proof: -
i s t o r s i o n free and t h e k-invariants of Given
g: L
+
?.
?
P
?,
A s i n t h e proof o f 5.3.1 t h e r e e x i s t s
N
P
Y
so t h a t
a r e of o r d e r power of
IT,(?)
p.
Suppose one has a commutative diagram
n
and a homotopy
$n-l:
L
-+
PHtn-l(?)
of
hn,n-lhng
N
r
f r e e , kn
of o r d e r
and
p
H"+l(M,Z)
p-torsion
in-lf.A s nY f r e e in-1l i f t s IT
is
A
to
If one considers
g.:
be l i f t e d t o a homotopy
w: L
-+
As j: M
-+
$: L - + P H t n ( f ) , 4: hng
-
w w
(iAf)
4n-l
can
f o r some
n).
K(T,(?). Hn(f,Z)
and consequently
K(IT~(?), n).
as w e l l and
as a f i b r a t i o n t h e homotopy
hn,n-l
4: hng
Replace N
A i
w w (iAf)
Hn(f,
IT^(?))
4= w
1
by
,
are s u r j e c t i v e
w -Gf,
.
w g:
which l i f t s
can be followed by a homotopy
n-1
H-maps i n t o
core1 lifting
*
w
hn ,n-1
-
(iAf)
%f
*
( % f ) = gnf
(and see remark 5.3.1.1).
Qn-l
P u t t i n g together 5.3.1,
181
B(n,p)
5.3.2,
t o obtain
Q : hng- i n f n
One can pass t o a limit
5.3.3 and 5.3.4 and choosing
B(n,p)
t o have p-torsion i n t e g r a l k-invariants one obtains:
5.3.5.
If H,(X,Z)
Corollary: (a)
u : [X, B(n,p)]
fb)
[X, B(n,p) lp
(c)
If f : X' f*:
(d)
+
+
is p-torsion f r e e then
Hn(X,
Hom(H,(X,Q)
-f
is surjective
Z P
, H,(B(n,p) , Q)) is
yields a surjection H*(f,Z)
X
[X, B(n,p)l
[XI, B(n,p)]
+
There exists a : X
-+
IIB(niy p )
injective
then
is a s u r j e c t i o n .
so that
H*(a,
i
zP
is s u r j e c t i v e .
5.4. H-maps i n t o B(n,p) W e study here conditions f o r H-maps
to
X
5.4.1.
-+
B(n,p).
We choose
Proposition:
Given a map
a: X
+
Let
K(Z n ) t o be H-liftable P' t o have i n t e g r a l p-torsion k-invariants.
B(n,p)
X, p
X
-+
be an H-space,
IIB(ni, p ) = B ( X )
so t h a t
H,(X,Z)
H*(a,Z)
torsion free. is s u r j e c t i v e .
i
There ex-ists an H-structure for
Proof:
H*(a,Z)
so that
a
is an H-map.
s u r j e c t i v e implies H*(a
surjective (as
B(X)
H*(X,Z)
A
a, Z):
H*(B(X)
is torsion free).
A
B(X), Z)
By 5.3.4
-f
H*(X
A
X, Z )
182
(a
BP
Non s t a b l e
a)*:
A
[B(X) A B(X), B(X)]
E im(a
HD(a, p ,
3.4.2.
[X
A
X, B(X)]
i s s u r j e c t i v e and
Now apply 1 . 5 . 1 ( a ) .
a)*.
Let
Proposition:
coker *;
A
-+
resolutions
X , 1~ be an H-space.
and
Suppose H * ( X , Z )
are p-torsion f r e e .
Then for any primitive element fx: X + B ( n , p )
80
x E Hn(X, Z ) P
there e x i s t s an H-map
that a [ f x ] = x.
5.4.2 follows from t h e more general proposition.
5.4.3.
N
H-map
Let
Proposition:
X,
u be as in
Given H-spaces and an
5.4.2.
N
+Y,
h: Y ,
so that n,(fiber h) are f r e e and suppose there
p0
e x i s t s a commutative diagram of H-spaces and maps:
-
the product muZtiplication.
t o an H-map
Proof:
N
f: X ,
p +
?,
Then any H-map
f: X ,
p +
Y,
uo
Efts
N
p.
One uses a Postnikov decomposition of
h
and induction:
H-maps i n t o
183
B(n,p)
a
Y
K( a,(fiber h ) ,
x
1
x-’
/
x
Pn
,n-1
N
fn-l
N
Y = 0
N
Given an H-map N
t o a map
f:
N
fn-yFn-l:
N
x, u
+
yn-y
N
N
un-1,
the l i f t i n g of
X + yn exists by o u r s t a n d a r d k - i n v a r i a n t argument.
N
u , vn) =
jwny
W
X
A
X
K(a ( f i b e r h ) , n )
/j n \
P2an
,,/ -
ln,n-1
N
‘n-1’
X
Fn-l
N
’n-1
K(an(fiber h ) , n )
fn-l
184
Non s t a b l e
TI: n
The homotopy ( c o r e 1 E,,,
E,).
Fn(?A x
jwn +
(We assume
BP
-
?A)
resolutions
covers
n
N
Fn-l*
Phn,n-l
TINN n
Fn-l
t o be a f i b r a t i o n and l i f t i n g s i n t o
hn,n-l
N
t o be exact l i f t i n g s not only up t o homotopy).
Yn
i s an H-map
p2an
As
1~*
+
A s coker *;
n
N
fn,
Hn(X
(hence,
A
X, Z ) 8 s n ( f i b e r h ) = H " ( X
coker
c*
-
8 1) i s p-torsion
isomorphism i w E i m ( u * Q 1) implies n
P-p w
a n ( f i b e r h ) ) = Hn(X,Z) Q a n ( f i b e r h )
Q 1: $ ( X ,
=
(r*Q l ) v ,
N
f' = jv
n
v: X
+ ?n
-+
n
Alter
furthermore, one has an H-structure
Fn
?A
to a lifting N
for
N
fn
H*(i,Z)
r*0 1,
HD(yn, 11, u n )
and e s s e n t i a l l y by 2.2.1 w
X, a n ( f i b e r h ) ) .
free and
w E im
K(an(fiber h ) , n ) .
A
+
so that
=
II
Ph
F n,n-l n
and N
N
Fn-l
( c o r e 1 E o , Em). This observation allows us t o pass t o a limit li.m(Yn,
'Fn) = Y, 'F: x
+ lim
f
Hn
N
= Y.
f
To apply 5.4.3 for t h e proof of 5.4.2 t a k e
The e x i s t e n c e of
a
Y = K(Zp.
follows from t h e observatoon t h a t
n),
H*(B(n,p), Q)
p r i m i t i v e l y generated and one has a 0-equivalence and an H-map $: B(n,p)
+
K(a,(B(n,p)).
necessary t o o b t a i n
$'I:
Replace B(n,p)
B(n,p) H Ip-p
by
B(n,p)(Cp),
K(*,(B(n,p))
= B(n,p).
and
$1 i f
is
H-maps i n t o
coker
The condition on
r’f i n 5.4.2
185
B(n,p)
and 5.4.3 can be recognized using t h e
following :
5.4.4.
, J, be a torsion f r e e connected graded
Let A,
Lemma:
algebra of f i n i t e type over coassociative.
h e n coker
if the dual of
A Q Z P
Proof:
Z.
Suppose
A
- - A 0 x) (J, : A -+
Hopf
is c o c o m t a t i v e and
i s p-torsion f r e e if and only
is a f r e e algebra.
F)J. and i m T c k e r ( F Q 1 - 1 Q F) . Cocomutativity implies i m F C ker(1-T). ( l - T ) k e r ( F @ 1 - 1 Q F) c k e r ( F @ 1 - 1 4 F), k e r ( F Q 1 - 1 4 F) = ker(1-T) n k e r ( F Q 1 - 1 Q F) @ W. A Q x = k e r ( F 4 1 - 1 Q F) @ V ( V = i m ( F 4 1 - 1 Q 8) C x Q Z Q which i s f r e e ) .
-A 0
J,
AS
i s coassociative
= [ker(l-T)
ker(FQ 1
-
(F
1)J. = (1 Q
18 F ) ] @ W @ V
and
im
F
As
lies in
t h e f i r s t summand., Coker
F
i s p-torsion f r e e i f and only i f
U = [ker(l-T) n k e r ( F @ 1 - 1 8 F ) ] / i m F i s p-torsion free. every f i e l d ker(TF 4 1
-
F
if
1Q
Now
ker(T Q 1
AF, 6 , , qF = A 0 F,
TF)/im TF =
U 4 F = [ker(l-T)
F = Z P
n
Q Cotor2’*(F,F) = 0
AF
U Q Q = 0 implies t h a t coker
- 18 i ) / i m
U
i s p-torsion f r e e .
6 0 F,
Cotory(F,F)
ker(TF Q 1
CotorE’*(Z,Z) $ Q F
i s t o r s i o n and
then
and
- 1 x TF)]/im -
i f and only i f
and for
= Q Cotor2’*(F,F).
J,F
%
U B Q = 0.
For
(+I* i s
a f r e e algebra.
= 0
i f and only i f
U 4 Z
P
186
Non s t a b l e BP
resolutions
A s a consequence of 5.4.2 and 5.4.4 one obtains:
5.4.5.
Corollary: Let
X, p
primitively generated and H,(X,
H*(X,Q)
commutative graded) algebra. by an H-mrp
f: X
Proof:
A = H*(X,Z)/torsion
H*(X,Z
As
Then every
P
) = A 8 Z
P
E*(u,Z)
5.4.5.1. H,(X,Z
P
E*(X,Z)/torsion 4 Z*(X,Z)/torsion
f r e e , and as
The non redundancy of t h e condition on t h e algebra
i n 5.4.5 can be seen by t h e following simple example:
$: K(Z,2)
-+ K ( Z p y 2 p ) .
f : K(Z,2) -+ B(2p,p)
H,(B(2p,p), vanish.
+
i s p-torsion f r e e and one can apply 5.4.2.
Remark: )
is
by 5.4.4 i f
r* i s p-torsion
Coker
Z ) a free (associative P x E PHn(X, Z can be r e a l i z e d P
i s coassociative and cocommutative and s o i s
p*: H*(X,Z)/torsion
coker
p-torsion free
i s a Hopf algebra and as A 8 Q
- then
H*(X,Z)
B(n,p) a[?’] = x.
-+
primitively generated A 0 Q A.
be an H-space
Zp)
This H-map cannot be l i f t e d t o an H-map
as t h e l a t t e r i s a mod p H-retract of
i s a f r e e algebra while i n
H,(K(Z,2),Z
P
1
BSU,
hence
p t h powers
P r o p e r t i e s of
5.5.
Examples:
Some p r o p e r t i e s of
w
zp)
and i f
0 # v~~ E PH
Zp),
H*(cP", 2n
(BUY Z
BU
B U ( 1 ) = CP" -+ BU
It i s c l s s s i c a l y known t h a t PH*(BU,
187
BU
induces an isomorphism pH2n+l(BU,Z) = 0
PH2n (BU, Z ) = Z ,
hence,
i n O # c E Z . then P v~~ = ~ ( ~ ) v ~ ~ + ~ ~ P( ~ - ~ )
P
A s a consequence one can e a s i l y see t h a t j
PH*(BU, Z P
5.5.1.1.
i s generated over
A(p)
by
PH2&' -2(BU,
@I
Zp).
a
O P C E Z . P
One has t h e following:
5.5.2.
Lemma:
Z )
H*(X,
P
We prove f o r
is injective, for
0
be an H-space,
PH a+2k(p-1)(f,
Proof:
mod(p-1).
X, p
is a free a l g e b r a and if H
then
n < p
Let
# H*(f.Zp)$(m;+,g
(as H*(X,Z ) P
-+
(f, Z ) P
BU
an H-map.
If
is i n j e c t i v e f o r some
is i n j e c t i v e f o r a l l k.
Z ) P
p = 2
p-odd.
m = k(p-l)+n,
H*(X,Z
As
2n
f: X
k < ko.
i s similar. Let
Suppose
PHh(f,Zp)
mo = k 0 (p-l)+n = mbp + m,:
i s a f r e e algebra H * ( f , Z ) v 1 ~ 1 1 )# 0 P 2(m0 m: 2m and H O(fyZp)v2mo# 0 . = P H * ( f , Z )v,
P
)
implies
-0
i s a f r e e a l g e b r a ) and
P<-'(f,
Z )v # 0. P 2m0
P
5.5.3.
Proposition ([Peterson]):
Proof:
BSU w
Il B(2n,p). n=2
m
Let
qo: BSU
Hence t h e k-invariants of
Ln
K(Z, 2n).
Put
B ~ U= B S U ( { ~ I ,$1.
n=2 BgU
a r e integral. of order power of
p.
188
Non s t a b l e
P Htpp(B^SU) = II K(Z,2n). n=2 and lift
P ll B(2n,p) n=2
B5U
Lift
BP
resolutions
P
II K ( Z , 2n)
+
to
a2: B^SU
-+
n=2
-+
P Ht2pp(BkJ) = ll K(Z,2n) n=2
a
P
II B(2n,p)
A BSU
1
to
al:
P ll B(2n,p) n=2
P ll B(2n,p) n=2
-+
BkJ.
a
n=2
i s a p-equivalence.
by 5.1.2 ( b ) a2a1
Looping twice one obtains
n=2 2
i s a homotopy equivalence through
As s2 a1
injective for
are injective. 2
s2 al,
composition
n < p.
kJ
applies t o
hence
n=2
88
By 5.5.2 (as o n l y
well)
8a28al
mod p
2p
2n 2 PH (a al,
Zp)
arguments a r e used it
2 PH*(Q2al, Z ) and consequently H*(Q a 1’ zp) P p-equivalence implies H*( $a,, zP s u r j e c t i v e ,
consequently a1 BSU +B$U
dim
and
a2
P
a _ ll B(2n,p).
axe p-equivalences and s o i s t h e
As BU w K(Z,2) x BSU,
n=2 P BU w ll B(2n,p) (B(2,p) P n=l P- 1 SU w II B(2n+l, p ) . n=l
5.5.4.
Corollary:
generated, aZgebras.
H,(X,Z)
Let
X,
FJ
K(Z,2)).
S i m i l a r l y one can argue t h a t
u be an H-space.
p-torsion f r e e and H*(X,Z
H*(X,Q) P
)
primitiveZy
and H,(X,Zp)
If x E PH2”(X, Z ) n < p then there e d s t s an H-map P
free
P r o p e r t i e s of
x E i m H*(f,Z
f: X +B(2n,p),
By 5.4.5 an H-map
Proof:
and H * ( f ,
P
exists.
f
189
BU
Z
injective.
P
By
BU w
P
B(2n,p)
and
n=l PH*(f, Z ) P
5.5.2
and consequently
Z ) P
H*(f,
are injective.
5.5.5.
Corollary:
F = Z
and 8 as Hopf algebras i f and only i f X, P
P
Let X,
be an H-space.
p
Then H*(X,F)
w H*(BU,
F)
i s H-p equivalent
m
, t o BU = B U ( { ~ I $1,
Proof: -
(q: BU wO_
II K ( Z , 2n)). n=l
One can obtain an H-map P
By 5.4.5
l i f t s t o an H-map
f
2P
i n j e c t i v e and as
5.5.6.
H*(X,Zp)
Corollary:
f : X +BU.
Z
w H*(BU,
IX,
pi f o r every
h e n X , M E GH(Bu) i f and o n l y i f
Proof: every H*(X,F)
p
H*(X,Z)
there e x i s t s
w H*(BU,Z)
f : X* P
p. W
e x i s t s an
If
f: X
This implies H*(BU,F)
H
for
H*(X,Z
P
H*(X,Z)
FJ
F = Z
or
p-equivalence
f
P
BU}.
P1 there
BU.
-$
X E GH(BU)
w
as a Hopf
Moreover, i n t h i s case f o r every f i n i t e s e t of primes
e&sts an H-map
is
kt
GH(BU) =
a7gebra.
Z ) P
H*(f,
i s a p-equivalence.
f
P
By 5.5.2
-
P'
X
)
FJ
H*(BU,Z
H*(BU,Z).
Q.
P
)
If
as a Hopf a l g e b r a f o r H*(X,Z)
w H*(BU,Z)
By 5.5.5 f o r every BU
and
p
there
X, p E GH(BU).
then
Non s t a b l e BP
resolutions
P1 i s a f i n i t e set of primes,
If
f : X + BU
P
P
(a =
Il
as
i s an
p)
is
BU
H-P1
Let
Corollary:
X
be a CW ecwpZex,
algebra on odd d i m e m i o m 2 generators.
equivalence.
k
and i f
GH(BU)
H*(X,
Z ) P
i s uncountable.
an exterior
If
2n+2k(p-l) + 1 rank &H (x,
is independent of
and HC
KP1
One can e a s i l y argue (and s e e 4.6.8) t h a t
5.5.7.
HA
zP)
= an
P2m: Qi-12m+1(X,Zp)
-+
Qi-Ianp+’(X,Zp)
injective
a ll B(2n+l, p ) n.
P-1 then
X W
n=l Proof: -
Z and H,(QX, Z are f r e e algebras, P P p r i m i t i v e l y generated and one can apply 5.5.4.
5.6.
H*(S2X,
Non s t a b l e BP
Adams r e s o l u t i o n s .
5.6.1. D e f i n i t i o n : Given an %spectrum E, = a strong E,
H*( X, Q)
A h s resolution f o r
X
{E
n’ Y n 1
and a space
X,
i s a sequence o f spaces and maps
Non s t a b l e
BP
Adam resolutions
19 1
so t h a t ( a ) hn:
Bn(x)
+
in-l(x)
(b)
If t h e f i b e r of
(c)
For every
i
i s a p r i n c i p a l f i b r a t i o n induced by a map
is
fn
and
n,
[e,(X), Ei]
::?i
l i m an = n-
connected then
a n
*
[X, Ei]
-.
is
s u r j e c t i v e and
5.6.1.1. spaces
Remark: X with
Note t h a t t h e c l a s s i c a l H,(X,Q)
K(Z P'
)
Adams r e s o l u t i o n f o r
# 0 does not satisfy 5.6.1 ( b ) but some weaker
property holds t o obtain an Adams s p e c t r a l sequence, hence, t h e r e s o l u t i o n i s s t i l l a good estimate f o r
X.
The e f f i c i e n c y of
spectrum f o r an Adams r e s o l u t i o n f o r
X
E,
as t h e underlying
i s measured by t h e r a t e of
Non s t a b l e
192
an
i n c r e a s e of with
H,(
,Z)
i n 5.6.1 ( b ) .
BP
resolutions
e.g.:
b e i n g p-torsion) t h e
monotonic non decreasing sequence
For p-torsion spaces ( i . e . :
,Z
K(
P
{an)
)
spaces
resolution yields a
-+ m.
{ B ( n , p ) ) do not form an a-spectrum as
B(n,p)
SB(n+l, p )
FJ
only f o r
But t h e s e r e l a t i o n s are enough t o e s t a b l i s h Adams r e s o l u t i o n s u s i n g {B(n,p)).
We s h a l l refer t o such a r e s o l u t i o n as a fnon s t a b l e )
BP
Adam8
resolution. We s h a l l o u t l i n e t h e c o n s t r u c t i o n o f t h e r e s o l u t i o n f o r spaces Z ) a free algebra. P c o n s t r u c t i o n can be obtained f o r an H-space X w i t h H,(X,Z)
with
H,(X,Z)
p-torsion f r e e and
H*(X,
X
A similar
p-torsion
free.
5.6.2. H*(X1,
Let
Proposition:
f : X1
-+
X
2'
Z ) / / i m f * are p-torsion f r e e ,
are f r e e algebras.
Let
Y = fiber f.
Suppose H*(Xi.
Z
Z)
H,(Xi,
P
and
Then: H,(Y,Z)
and
H*(X1,
Zp)//im
i s p-torsion f r e e
and the EiZenberg-Moore s p e c t r a l sequence: H*(X2, Tor
collapses f o r F = Z
P
generated by
F) (H*(X1,
and Q.
k e r &H*(f,F)
F ) , F) ==> H*(Y,F)
Moreover, i f
then f o r F = Z
P
M2
is the f r e e algebra
or Q
M H*(Y,F)
as an H*(%,
F)
FJ
module.
f*
[H*(X1, F)//im H*(f,F)] Q Tor 2(F,F)
Non s t a b l e
Proof:
BP
Adams r e s o l u t i o n s
One can e a s i l y argue t h a t under t h e above hypothesis
W*(Xi, Z )
a r e p-torsion f r e e and i f
by
Z)/torsion
QH*(Xi,
Furthermore,
then
Ai
i s t h e f r e e Z-algebra generated
Ai 0 F wH*(Xi,
F = Z
F)
induces a homomorphism A( f ) : A2
f
A ( f ) Q F = H*(f,F). so that
193
P
-+
(Actually one can replace ;Xi,
Ai = H*(Xi,
Z)/torsion.
A1
SO
and
Q.
that
by p-equivalents
f
This i s done by s t a r t i n g with a
commutative diagram
then use 4.3.1 t o obtain
2.1
are t h e d e s i r e d spaces).
= Xi({p>, lli)
A1 = N1 Q N 2 , Mi
, Ni
i m QA(f ) , N1 w M1. As
rank(Mi Q F ) ,
generated by
A2 = M1 0 M2,
k e r QA(f), N1
Tor
rank(Ni Q F)
Tor
by
M2 Q F
A2 Q F Then
A2 Q F
t h e same holds f o r
M2
f r e e algebras,
Now, one has
(A1 0 F, F) w (N2 Q F) Q Tor a r e independent of
(4B F,
F).
F
( F = ZP
(F,F).
or 61)
Non stable BP resolutions
194
It suffices to show that the Eilenberg-Moore spectral sequence FEtw collapses for F = Q. For then, using a Bockstein spectral sequence argument to obtain rank Hn(Y, Zp) 2 rank Hn(Y,Q) rank Hn(YaQ) =
C
rank mE!
C
rank Tor
q+q'=n
999'
and one has: =
C
q+q'=n
rank
%'9,q' =
H* ( x2 ,Q) =
q+q ' =n
9a9'
(H*(X,,
Q), Q)
=
= rank Hn(Y, Z ) 2 rank Hn(Y,Q)
P
2
Hence all the (weak) inequalities are actually equalities and
'E2
=
Z 'Em.
Moreover, rank Hn(Y, Z ) = rank H"(Y,Q) implies the collapse of the P Z - Bockstein s.s and H*(Y,Z) is p-torsion free.
P
The collapse of the Eilenberg-Moore spectral sequence f o r can see by comparing
F = Q one
Non s t a b l e
k(Ni)
where
'"Mi
= K(QIVi))
BP
H*(?(Ni,
195
Adam r e s o l u t i o n s
Z),
Z ) / t o r s i o n = Ni
and s i m i l a r l y f o r
1.
k(M2) observe
fiber k ( A ( f ) ) = K(N2) x
To s e e t h a t
The c o l l a p s e of t h e s.s f o r t h e r i g h t hand s i d e f i b r a t i o n o f 5.6.2.1 i s obvious
.
I f one a p p l i e s 5.6.2 f o r
5.6.3.
x1
= w
one g e t s :
Corollary:
(a) Let
X
be a space with H,(X,Z)
a f r e e algebra. F) = Tor
H*(QX,
Ib) I f
X1s
X2s
H*(Y,F)
as an j:
Qx2
Then H*(flX, H"
f, Y
N
Of
P
i s p-torsion f r e e and
z
P
or
Q) as coazgebras.
Q i r n H*(j ,F)
H* X ,F)//im H * ( f , Z ) 1
module and
from the Puppe sequence
Y
Z )
are as i n 5.6.2 then
F)//im H*(f,Z)
Q5
Z)
' y F ) ( F , ~ ) , (F =
H*(r, -+
p-torsion free and H*(X,
ax2 -L Y
-x1
f
x2.
i m H*( j ,F)
comoduZe,
Non s t a b l e BP r e s o l u t i o n s
If i n addition H*(C2X2, F) is a f r e e algebra (always holds for Q) then the above representation of
F =
i s an
H*(Y,F)
algebra representation.
5.6.4.
X be a space,
Let
free algebra. Let
W e construct a BP Adams r e s o l u t i o n f o r
fo: X
-+
Z ) a P as follows:
p-torsion free and
H*(X,Z)
EO(X) = GB(ni, p )
H*(X,
X
Zp)
yield a surjection H*(fO,
(and
1
obviously of
el:
EO(X)
so t h a t
+
H*(fo.
K(A)
klfO
-
Indeed,
to
C(fo)
-+
Let
kl:
EO(X)
f a c t o r s through H*(C(fO), Z )
-
B.
Lift
can be done as Lift
-+
F(K(A)) which induces
H*(E,(X),
fn
to
en
to
fn+l: X
obtained by 5.6.2,
-t
t h e cone on
kl
C(fo)
M
exists, Mn
H*(X,
-+
kn:
H*(En+l(X)y
Z)
K(A)
is p-surjective).
Let
The cohomology of
to
C(fo).
H*(En(X),
-
En(X)
-+
K(&M,),
knfn
En+,(X)
E
n+l
-*
(which
= f i b e r kn. (X)
(as H*(F(K(QMn)) = p ( n i , p ) , Z p )
i s p-torsion f r e e and
Z)
a f r e e algebra,
so that
F(K(&M,))
1
algebra).
-+
f a c t o r s through
i (n
fi
H*(fo,Z)
As
fo.
i s p - i n j e c t i v e and
Z)
Z ) 4 Mn, P k e r H*(fny Z p ) . Map
En+l(X).
5.6.3.
H*(EO(X),
k 1y
kn: En(X)
H*(fn, Z )
-
One can l i f t
Ei(X),
Zp)
generating t h e i d e a l
cnfn
can be chosen
FK(A) = llB(mj, p ) , kl
C(fo)
i s p-torsion f r e e .
Suppose a r e s o l u t i o n
Mn
-+
Lift
Y.
el
p-torsion f r e e ,
zP).
A = ker QH*(fo,
j
i s a p-surjection
H*(C(fO), Z )
Q)).
can be is a free
Non s t a b l e
BP
Adams r e s o l u t i o n s
F i n a l l y w e s h a l l o u t l i n e t h e construction of a f o r an H-space
X
generated and
H,(X,
with
p-torsion f r e e ,
H,(X,Z)
Z ) P
197
BP
Adam r e s o l u t i o n primitively
H*(X,Q)
a f r e e algebra:
One s t a r t s with an H-map fo: X
B(nj, p) = EO(X),
Il
-+
H * ( f o y Z) p-surjection
j €Jo
X -+K(PH*(X,Z)/torsion) = KO
This i s done by l i f t i n g
fo:
X
m(BKO) =
-+
B(nj, p )
Il
(BKO
-
t o an H-map
t h e c l a s s i f y i n g space of
KO
XJ0 QBKO = K O ) .
H*(foy Z )
Let
i s p-surjective.
=
E O ( X ) -+$,(X)
B(nj, p ) ,
ll
Realize klfO
J€Jl
x
t o an H - ~ P fl:
fo
is in
im[X
A
-+
X, B1(X)]
Z ) P
A c H*(EO(X),
k e r H*( f o y Z p ) .
algebra generating kl:
Z) i s a f r e e a l g e b r a one can s e e t h a t
H,(i2F(BKO),
AS
A
- *.
be a sub-Hopf
by an H-map
E ( X ) = f i b e r kl 1
Let
G1(x)
E ~ ( x ) . Indeed, if B ~ ( x )=
-+
[X
A
X, E1(X)]
X
and as
A
X
-+
lift
then
E1(X)
E1(X)
A
y i e l d s a mod p s u r j e c t i o n o f i n t e g r a l cohomology by 5.3.5 ( c ) [E1(X)
A
B1(X)]
E1(X),
the multiplication i n that
A1
M
fl
+
i s an H-map and
Z ) P
-+
A
X, B1(X)]
by a map
E1(X)
i s an H-map.
im[H*(E1(X),
[X
i s s u r j e c t i v e and one can a l t e r E1(X)
B1(X)
-+
E1(X)
as Hopf algebras, as
X
+
H*(E1(X),
Z p ) w H*(X,
H*(B1(X),
Zp)]
E (X) 1
E1(X)
A
a f r e e algebra and. one can proceed by r e a l i z i n g
A1
E1(X)
+
B2(X) w
-+
TI B(nj, p ) e t c . J€J*
-+
so
Z ) Q A1, P
is central.
j,:
B1(X)
A
1
M
k e r H*(fly
by
E1(X)
2 )
P
is
Non stable BP r e s o l u t i o n s
5.7.
Some simple a p p l i c a t i o n s . The s i m p l i c i t y r e f e r r e d t o i n t h e t i t l e of t h i s s e c t i o n has t o do
only with t h e f a c t t h a t w e use mostly 3-stage computations.
BP
r e s o l u t i o n s i n our
The a r i t h m e t i c a l c a l c u l a t i o n s however m a y be complex.
5.7.1. Proposition:
Let
be an Ho
X
space,
p-torsion free.
H,(X,Z)
n +1 If 'nn
#
(X)
for
= 0
'nn(X)
0,
n < n
0
or
n -2p+3 &H O (x,
Proof:
To:
0
(x, zp) #
0
0
n0-2p+3
zp) #
Map
then either &H
0
X
#
(X,Q).
QH
BO(X) =
-+
so that
II B(ni, p )
H*(fO, Z ) P
i EJ
is
We may assume t h a t f o i s e f f i c i e n t i n t h e sense t h a t t h e n. n n image of t h e composition H '(B(ni, p ) , Z p ) -+ H ' ( B 0 ( X ) , Z p ) -+ H i ( X , Z p )
p-surjective.
n. H '(
i s not i n t h e image of
Il B(nJ,p),Zp)
-+
n H i(Bo(X),Zp)
-+
n H i(X,Z).
j
Note t h a t
QHm(B(ni, p ) , Z p ) = 0
m :n
unless
i
(2p-2) and t h a t
ni +2k(p-1) (B(ni, p ) , Z ) # 0 f o r all k 2 0. L e t mo be t h e minimal &H P m m 0 i n t e g e r with k e r H ( f o , Z p ) # 0. Let A = k e r H O ( f 0 , Z ) / t o r s i o n and let
il: B O ( X )
+K(A, m )
m QH
'(klY
fl: X
-+
'nn(
1
fl
Q ) # 0. E1(X)
n < mO-1.
realize
0
-
ilfo
and
Z ) P
H*(fly
i s an isomorphism f o r
As
1
n < m
0'
n'
= A
o z
P
:
space
i,
fo
then
Obviously
dim
lifts to
5 mo,
'nn(E1(X))
(E ( X ) ) # 0, m -1 1 0
o zp
i s an Ho
i s an isomorphism i n
We s h a l l show now t h a t
(E ( X I ) Pn mO-1 1
X
E (X) = fiber
If
it.
A.
hence = 0
for
as a m a t t e r of f a c t
199
Applications
kl
obviously f a c t o r s as
BO(X)
II
-+
k1
B(ni,P)
K ( A , mo)
5
n <m
i- 0
i
Now
ll B(ni, p ) n.=m 1 0
zero map on
Z
d i v i s i b l e by
1
n
5
n <m i- 0
K(A, m ) 0
k1
B(niy P )
cohomology by t h e e f f i c i e n c y of
P p
k i is 11
Hence
fo.
Now, f o r any map
n m (klil) Q Zp = 0.
and
must y i e l d t h e
0
k i i ) : B(ni, p )
K ( A , mo),
-+
< mo
n
if
n m (kl( i1 ) 0 Zp = 0
then
i
Hurewicz homomorphism a n : nn(B(ni, p ) ) satisfies u
n
8 Z
minimality o f
n
P
= 0
-
m
5.2.2 ( e ) ) .
(kl) 8 Zp = n 0
m
n
mo-1 j
Q Zp @ (
(klil)
C
n
n.<m m~ 1 0
(klil
=
m ‘(k,,
i E J
m o - ~ + 2 ( x ,zp) # QH
QH
m 0 QH (klil,
mo
(2p-21,
-
(X, Q).
Z ) = 0 P
implies
QH
5.7.2.
Let
x
ni
ni
H*( X, Zp)
# 0,
( II B(niy P ) , Q ) iEJ
m0’
If for some
m
Proposition:
j E J
but then
mo 2p +2
#
n +1= m0’ 0
Suppose
0
Q ) # 0 i m p l i e s QH
F i n a l l y observe t h a t
generated,
Now, i f for no
m
*I,QH
0
( k ( i ) ) 8 Z ) = 0. l P
E ( X ) = coker n ( k l ) , mO-1 1 m 0
consequently f o r some
as
Hence,
(E ( X ) ) 0 Zp = coker n (kl) 8 Z = A 0 Z 1 m0 P P‘
= no,
n > ni
Hn(B(niy p ) , Z )
0
Now one can e a s i l y s e e t h a t
n
-+
k-invariants w i l l v i o l a t e t h e
(as 0-mod p
B(ni, p )
as t h e
0
-
X, p
0
(B(moy p ) , Zp)
be an H-space,
j
E J
,QH
# o
H*(X,Q)
n
j
= m
o
m O(X,
zp’.
primitively
exterior algebra on odd dimensional generators.
is 2n connected
mod p
(i.e.
H2n+l
(x, zP) #
0
and
Non s t a b l e
200
(XI
Ht
P-1
a(n,p)
Proof:
resolutions
in
PH*(X, Z
i s a two s t a g e
X
a(n,p) + 2
dim
Put
fo:
X
-+
satisfy:
P
is a
Zp[P1]
( X , Zp)
by
fo,j: X
B O ( X ) = JIB(2nj+lY p ) .
BP
resolution.
One can
as follows:
module,
2n .+l a j E PH
every
xi
where
i=l
I n t h i s range
PH*(X, 2 P
resolve
xi,
II
w
BP
+
P1 E A(p)
B(2n +1, p ) , f o Y j - H-mp.
J
Note t h a t
J
i n dim
2
Let
a ( n , p ) + 2.
B(2nj+l, p) + K ( Z
kl,j:
PY
2n
'
+ 2rj(p-l)),
+ 2 r . ( p - l ) + 1) products t a k e n over d l J
N
mod(p-1).
= klyi:
Lift
J
BOyi ( X )
-+
Ki
t o an H-map
j
with
nj
5
i
201
Appli c a t i o n s
2i+1+2k( p-1) f o y Z p ) [PH
one can choose
fo
lifts t o
put
k
so that
l,i
fl: X
xi = Hta(n,p)
-+
i
- *.
klyif0
X ) , Z ) ] = PH2i+1+2k(p-l) (X’ZP) P Put
Ei(X) = f i b e r k l,i
which i s a p-equivalence i n
IIE.(X) 1 i
( E (X))
(Bo,i
Ht
and then
a(n,p)
( X ) mp IIXi
dim
5
then
a(n,p).
i s the
d e s i r e d decomposition,
To i l l u s t r a t e t h e use of t h e
BP
r e s o l u t i o n s for s p e c i f i c
computations we have:
5.7.3.
Proposition:
PlTm(s3)
If 4 5 m < 2p2
=
0
and
#
kl
i f
m = a-1
mod 2p-2
otherwise
Consider t h e r e s o l u t i o n :
have t h e p r o p e r t i e s :
H*(koyZp) (1 8 u4p-1)
(ui E PH1(B(i,p) ki
then
m f 1, 2
Z
ko
2
o if 0
Proof: -
-
, Zp)
-
H*(koy Z p ) ( ~ a + l
1
0 1) = P u3-
E PH4p-1(B(3y~).Zp)y H*(klyZp)u4p-l
a generator).
give r i s e t o t h e following tower:
k o y kl
are H-maps and
= $u2p+l klkO
-
@
0.
1
Non stable BP resolutions
202
E2 = fiber i(. 1
s3
I
sljl = Qkl. is an isomorphism for s 5 2p2-2 hence
HS(h 2,Z P )
isomorphism for m
2
2p -2.
2
To compute coker(a,(k B(3,p)
))
m
(h ) 2
is an
It can be seen that
m $ 1, 2, (2~-2)
0
m
T'
'[coker
T
'[coker
TI
m+l
(kO)]
m
2
(2p-2)
(k ) ]
m :1
(2p-2)
m+2 1
consider k B(3,p) 0,2:
-f
B(4p-1).
is a mod p retract of SU, PH*(kOy2, Zp )u4p-1 =
W
By 5.5.3, 4p-1 # 0 and
Z ) is an isomorphism in dim by &H*(kOy2, Z ) = PH*(k P 02' P 2 2 4p-1 5 b 5 2p + 2p-1. Hence for 2p-1 5 s 5 p + p-1 , s = 1 + k(p-1) , 2 2 k 5 p, one has
Applications
1
U
and degree
'II
2s+l
203
I
2s+l
u2s+1
degree i2s+l] * c,
= [degree
(k 0,2
c E Z,
(C,P) = 1. By a theorem of Bott (and as degree
a
2s+l
degree(a2s+l) Degree n
Now
2s+l
= s!
((k-l)p + p
retract of
SU)
and by the Brown Peterson c o n s t r u c t i o n
. cl,
k-2 = P (k ) 0,2
i s a mod p
B(3,p)
(c,,p)
= 1 and
k
i s given by
s - 1 = k(p-1)
i s then
- k+l)!
= c3(k-l)!p
k-1
1< k
p-1.
Hence,
degree ~ ~ ~ + ~= c( * pk , ~'[coker , ~ )~ ~ ~ + ~ (= kZ ~ , ~ ) 1 P'
To compute degree n2s+l(k0,1)y
isomorphism f o r
a).Hence,
m = 4p-1
H*(ko,ll
(as no
kOs1: B(3,p)
k
2p+l
B(2p-1, p )
I k 5 4p-1
ZP) = €I*( fi-2koy2),
consider
i s d i v i s i b l e by
,
204
Non s t a b l e
degree *2s+l(k0,1)
= degree n
2s+2p-1
BP
resolutions
= pec
(k 0,2
s = 1 + k(p-1)
for
~ ( klyl: k ~ B ()2 p + l , p ) + B ( 4 p - l Y p ) , 1 2 k 2 P-1. To compute ~ ~ ~ + put
.i
= kl
klYl
H*(klYly Zp)u4p-1 = u ~ + ~hence, ,
19
isomorphism for
5
4p-1
i
2 2p2
.
Thus, for
2
QHi(klyl,
Zp)
i s an
2 k 5 p-1
I"' Degree u
k
= c
pk-l,
1
degree .n,(k
091
)*degree a,(k
Consequently; degree II Hence,
'[coker
'[coker
71
2s+l
71
k-2 = c2p
6
degree
2s+l
2s+l
(k 1,2
( k ) ] = Zp 0
(k ) ] = 2 1 P
for
s
,
hence,
) + degree a,(k ) * d e g r e e n,(k 1,1 OY2 L2 = c'sp for
(k1,2:
s 'p,
2p-1
B(4p-1, p ) +B(4p-1, p s
3
1, mod (p-1)
s I 1 (p-1
.
and
5.7.3 follows.
A more t e c h n i c a l computation b u t along t h e same l i n e s leads t o
5.7.4.
Proposition:
Define
j , k, a,
-
O z a < p.
!&en
Given i n t e g e r s n , m, n 2 p-1, by
m = j
+
k(p-l),
0
5
j < p-1,
m 5 p(p-1) n+j = ap
- 1. +
a,
= o
1.
Applications
where
Y (b) i s the highest p m e r o f p dividing b. P [Krishnarao] 1.
where
(Compare t h i s A t h
1,2
Outline of computations: The Adams resolution in this range has just
3 stages and is given by:
SU(n)
fo
P-2 B
0
= B' x B1, 0
BA =
n
B1 =
j=o
n+j = a(j)p +
;(j),
kg
,
BO
B1'
P-2 B ( a j , p),
B ( Y j , p),
aj = 2n+2j+lY
j=o
-.. Yj = 2n+2j+1+2(p-l)(Z(:(j)+l) where a(j) by
su
is defined for every integer j
0 L c ( j ) < p. kOy1 = plko: SU + B A
is given by
a-
PH '(kOyl, Zp) is an isomorphism, 0 5 j 5 p-2, k0,2 = p2ko: SU +B1 given by PHv j (ko,2, Z ) P
is an isomorphism 0 5 j 5 p-2, kl is given by
YA klil: B + B satisfies PH j(kOy2,Z )u =pa(')+lu 0 1 P Yj klko N
y.
A s in
is
5.7.3 here too we have
aj'
ki H-maps and
Non s t a b l e
206
n'
2n+2s
BP
(SU(n)) = '[coker n
and t o compute these one has t o compute J = 1,2.
degree n,(k
2(p-E1)
times t o obtain maps
, ZP )
PH*(
resolutions
031
)
2n+2s+l(k0 11
degree n,(k
i,j
i s camputed by delooping
8U
+
T
i = 0,1,
SU + B ( a j , p )
B(2n+2J+l + 2(p-z1), p)
an isomorphism i n t h e range of i n t e r e s t .
'[coker
)
2n+2m+l(k0,l 11 = Z rl(n,m) P
with
One obtains
,
A d i r e c t computation y i e l d s
'[coker
n
2n+2m+l(k0,2 11 = Z r2(n,m) P
r2(n,m) = 0
klkO
-
w
if
k
28,
yields for
d = degree n 2n+2m+l(k1 , 2 1: P hence,
Y d = Y (a+k).p. P P
rl(n,m)
- = . pa+l
5.7.4 now follows.
As a f i n a l application of t h e BP resolutions we have:
r2(n,m) Yp(d) 9
Applications
5.7.5.
bundle 5
o
= H 4n+l ( y , z )
dim
>_ 4n.
of (complex) dimemion 2n+l
Then f o r any complex vector
the obstruction f o r 5
4-sections are the 4 top integral &ern classes of
Proof:
Suppose
i s 2 torsion free i n
and H*(Y,z)
and 3 torsion free i n
411-2
dim
Given a 4n+2 dimensional space Y.
Proposition:
H4n-l(y,z) =
207
t o hose
5.
The obstructions i n question axe t h e obstructions i n t h e
l i f t i n g problem:
BU( 2n-3)
(where
f
5).
classifies
To study t h e l i f t i n g problem one prime at a t i m e note t h a t i f i = l,..
hi
.,k,
- a mod Pi
p a r t i t i o n of if
axe spaces s o t h a t f o r every
equivalence i n
P, then
f lifts to
f?:
Let
P1 = I21,
Let
X3
Y
f: Y -+
xi
Then h
5 4n+2
{P.1
and
BU(2n+l) l i f t s t o
i .
f o r every
P2 = (31,
be t h e f i b e r of
t h e Chern classes.
dim
..,k
i = 1,.
BU(2n-3)
is a
i f and only
i.
P3 = P
-
{2,3].
w: BU(2n+l) factors
i’
f a c t o r s as
h
i
X
3
-+
n K(Z, j=o
BU(2n-3)
FJP <, 4:+2
-
4n-4+2j) 1
X3
realizing
-+
BU(2n+l)
208
Non s t a b l e
and t h e o b s t r u c t i o n s o f an
H
4n-4
n
i s odd,
3
l i f t i n g are obviously t h e Chern c l a s s e s .
consider
(ko, Z) = w 4n-4,
surjective.
f
re.solutions
X1 we d i s t i n g u i s h between two cases.
To c o n s t r u c t If
N
BP
koh
- *.
Such a
is surjective i n
H*(ko, Z 2 )
e x i s t s as
ko
dim < 4n+4.
H*(h,Z) kl
is
i s given by A
H*(kl,Z2)u4n+2
i s homotopic t o through
= Sp4’2~4n-4. B(4n-5,2) = QB(4n-4, 2 )
Ql.
d i m 4n+4,
By 5.3.1
hence
h2
klhl is a
- *. mod 2
d i m 4n+2.
For n
even t h e r e s o l u t i o n i s given by
+
El-
kl
B(4n+l, 2 )
QH*(klyZ2) i s a monomorphism equivalence through
Applications
As H*(Y,Z)
is 2-torsion free in d i m > 4n-4 kof
209
-*
if and only
if H*(kof, Q) = 0. As imH 4n-4+2J(ko,Q) are the Cherm classes f lifts 4n-1 4n+l to El and as by hypothesis H (Y,Z) = 0 = H (Y,Z) f lifts further to
"f:
Y+ X 1'
Similar arguments hold for 2n
1
W4n-2a
2n
2
5.
Here we have three cases:
mod 3
imH4n-2a (ko, Z)
mod 3
a = 1, 2.
210
Non stable BP resolutions
211
Bibliography
Adams, J.F.
,
1. On t h e non existence of elements of Hopf invariant one.
Ann. of Math. 72 (1960), 20-104.
2. The sphere considered as an H-space Quart. J. Math.
3.
, Oxford
mod p.
(2) 12 (1961), 52-60.
H-spaces with f e w c e l l s .
Topology 1 (1962), 67-72.
Araki, S. and Kudo T., Topology of
Hn spaces and H-squaring operations. Mem. Fac. Sci. Kyusyu Univ. , 10 (19561, 85-120.
Arkowitz, M.,
Homotopy products f o r H-spaces.
Mich. Math. J. 1 0 (1963),
1-9. A r k d t z , M. and Curjel, C.R., 1. On t h e number of multiplications of an H-space.
Topology 2
(1963)3 205-209. 2. Homotopy commutators of f i n i t e order, 11. Quart. J . Math., Oxford (2) 1 5 (19641, 316-326.
3. On maps of H-spaces.
4.
Topology
6 (19671, 137-148.
Some properties of t h e e x o t i c multiplications on the t h r e e sphere.
Quart. J. Math., Oxford (2) 20 (1969), 171-176.
Berstein, I . and Ganea, T . , Homotopical nilpotency.
I l l . J . of Math. 5
(1961), 99-130. Bott, R . , 1. The space of loops on a L i e group.
Micpigan Math. J.
5 (19581,
35-61. 2. The stable homotopy of t h e c l a s s i c a l groups. (1959)s 313-337.
Ann. of Math. 70
212
Bibliography
Browder , W.
,
1. Torsion i n H-spaces.
Ann. of Math. 74 (1961), 24-51.
2.
Homotopy commutative H-spaces.
3.
Higher t o r s i o n i n H-spaces.
Ann. of Math. 75 (1962), 283-311. Trans. h e r . Math. SOC. 108 (1.963)~
353-375.
4.
Trans. Amer. Math. SOC. l o 7
On d i f f e r e n t i a l Hopf algebras. (1963)s 153-176.
5.
Homology r i n g s of groups.
Brown, E.H.
and Peterson, F.P.,
A spectrum whose
a l g e b r a of reduced p t h powers. Copeland, A.H. 1.
h e r . J. o f Math. 90 (1968), 318-333. 2
P
cohomology i s t h e
Topology 5 (19661, 149-154.
Jr.,
Binary operations on s e t of mapping c l a s s e s .
Mich. Math. J.
6
(1959) 3 7-23. 2.
On H-spaces with two n o n t r i v i a l homotopy groups.
Proc. Amer.
Math. SOC. 8 (1957), 184-191. Croon, F.H.,
Homotopy t y p e of t o t a l loop spaces.
Portugal Math. 30
(1971) , 11-16. Curjel, C.R.,
On t h e H-space s t r u c t u r e s o f f i n i t e complexes.
Comment.
Math. Helv. 43 (1968), 1-17. Curtis, M. 1.
,
F i n i t e dimensional H-spaces.
B u l l . Amer. Math. SOC. 77 (1971),
1-12 and 1120. 2.
H-spaces
mod p . 11.
H-spaces, Neuchatel A&t
1970.
Lecture
notes i n Math. 196, Springer, 11-19.
C u r t i s , M. and M i s l i n G.
, B u l l . Amer. Math. SOC. 76 (1970), 851-852.
1.
Two new H-spaces.
2.
H-spaces which are bundles over
(1971) 9 27-40
-
S7.
J . Pure and Appl. Algebra 1
213
Bibliography
Freyd, P., 1.
S t a b l e homotopy.
Proceedings o f t h e conference on c a t e g o r i c a l
a l g e b r a ( L a J o l l a 1965) , Springer (1966). 2.
S t a b l e homotopy 11. AMS Proceedings o f symposia i n pure mathematics , Vol. XVII (1970) , 161-183.
Fuchs, M . ,
[ ,Y] and loop f i b r a t i o n s I.
The functor
Mich. J. of Math.
1 4 (1967), 283-287. G i l b e r t , W.J., Homotopical n i l p o t e n c e of t h e seven sphere.
Proc. h e r .
Math. SOC. 32 (1972), 621-622. Harper, J . R .
,
1. Homotopy groups of f i n i t e H-spaces.
Comment. Math. Helv. 47
(1972) 3 311-331. 2.
mod 3
The
homotopy t y p e of
F4.
Localization i n group theory
and homotopy theory ( S e a t t l e 1974), Lecture notes i n Math. 418, Springer (1974).
3.
58-67.
H-spaces with t o r s i o n .
Harrison, J.
, Scheerer
H.
, Zur
(Mimeographed). homotopietheorie von abbildungen i n
homotopieassoziative H-ram.
Archiv der Mathematik Vol. 23 (19721,
319- 32 3. H i l t o n , P . , Homotopy theory and d u a l i t y , Gordon and Breach Science Publishers
, 1965.
Hilton P . , Mislin G. and Roitberg, J . , Localization of n i l p o t e n t groups and spaces.
North Holland Mathematics Studies 1 5 (1975).
Hilton P. and Roitberg, J . , On p r i n c i p a l
S3
bundles over spheres.
of Math. 90 (1969), 91-107. Hu. S . , Homotopy theory, Academic P r e s s , NY, 1959.
Ann.
Bibliography
214
Hubbuck, J . R . , 1. Some r e s u l t s i n t h e theory of H-spaces.
Bull. Amer. Math. SOC.
74 (1968)s 965-967. 2.
3.
Topology 8 (1969), 119-126.
On homotopy commutative H-spaces.
Generalized cohomology operations and H-spaces of l o w rank. Trans. Amer. Math. SOC. 1 4 1 (19691, 335-360.
4.
F i n i t e l y generated cohomology Hopf algebras.
Topology 9 (1970) ,
205-210.
James, I . M . , 1. Reduced product spaces. 2.
Ann. of Math. ( 2 ) 62 (1955), 170-197.
Multiplications on spheres 11. Trans. h e r . Math. SOC. 84
(1957) y 545-5580 3.
On H-spaces and t h e i r homotopy groups, Quart. J. Math.
, Oxford
Ser ( 2 ) 11 (1960), 161-179. Krishnarao, G.V.
,
1. Unstable homotopy of
90-97 2.
O(n).
Trans. h e r . Math. SOC. 127 (19671,
*
Nonstable homotopy groups of
U(n).
t o modern Analysis and Algebra.
General topology r e l a t i o n s
Proc. Kampur Topology conference
1968, 173-174 (1971). Kumpel, P.G., 1.
Jr.
L i e groups and product of spheres.
Proc. Amer. Math. SOC. (19651,
1350-1356. 2.
On p-equivalences of
mod p
H-spaces.
Quart. J . Math.,
Oxford 11, 23 (19721, 173-178. Larmore, L.L.
and Thomas, E . , Mapping i n t o loop spaces and c e n t r a l group
extensions.
Math. 2. 128 (1972), 277-296.
Bibliography
215
Lin, J.P., 1. H-spaces With f i n i t e l y generated cohomology algebras.
Bull. Amer.
Math. SOC. 80 (1974), 1233-1230. 2.
Torsion i n H-spaces I . , Ann. of Math.
Liulevicious, A.,
( t o appear).
The f a c t o r i z a t i o n of c y c l i c reduced powers by secondary
cohomology operations.
Memoirs AMS 42 (1962).
Massey, W.S., Peterson, F.P., The mod 2 cohomology s t r u c t u r e of c e r t a i n f i b e r spaces.
Memoirs of t h e Amer. Math. SOC. 74 (1967).
MW, P. and Zabrodsky, A., H*(Spin ( n ) ) as a Hopf algebra.
(To appear
Milgram, J.R., "he bar construction and abelian H-spaces, Ill. J. Math. 11 (1967), 242-250. Milnor, J . and Moore J . C . ,
On the s t r u c t u r e of Hopf algebras.
Ann. of
Math. 81 (1965), 211-264.
Mimura, M . , 1. Quelques groupes d'homotopie metastable des espaces symmetriques
Sp(n) e t 2.
U(2n)/Sp(n).
C.R.
Acad. Sci. P a r i s 262 (1966), 20-21.
The homotopy groups of L i e groups of l o w rank.
J. Math.,
w o t o Univ. 6 (19671, 131-176. Mimura, M., Serre.
Oneill, R.C.
and Toda, H . ,
On p-equivalence i n t h e sense of
Japan J. Math. 40 (1971), 1-10,
M i m u r a , M. and Toda H . 1. Homotopy groups of
SU( 3), SU(4), S p ( 2 ) .
J. Math. Q o t o Univ.
3 (1963/64), 217-2502.
On p-equivalences and p-universal spaces. Comment. Math. Helv. 46
(1971)9 87-97
-
216
Bibliography
Mislin, G.
,
1. H-spaces mod p
I.
H-spaces, Neuchatel Aout 1970.
Lecture notes
i n Math. 196, Springer (1970), 5-10. 2.
The genus of an H-space.
Battelle
Symp. on Algebraic Topology.
S e a t t l e 1971, Lecture Notes i n Math. 249, Springer (1971), 75-83. N a y l o r , C.M.
, Multiplications
on
SO(3),
Mich. Math. J . 1 3 (1966), 27-31.
Porter, G.J.,
1. Homotopy nilpotency of
S
3
.
Proc. Amer. Math. SOC. 1 5 (19641,
681-682. 2.
On H-fibrations.
Quart. J. Math., Oxford
Peterson, F., The mod p homotopy type of Mexicana ( 2 )
BSO
and
2 ) 22 (1971), 23-31. F/PL.
Bol. SOC. Math.
1 4 (19691, 22-27.
S i g r i s t , F., Groupes d'homotopie des v a r i k t d s de S t e f e l complexes. Comment. Math. Helv. 43 (19681, 121-131. Smith, L . , Lectures on t h e Eilenberg Moore s p e c t r a l sequence.
Lecture
notes i n Mathematics 134, Springer (1970). Spanier, E.H.
, Algebraic
Topology, McGraw H i l l .
S t a s h e f f , J. ,
1.
On extension of H-spaces, Trans. h e r . Math. SOC. l o 5 ( 1 9 6 2 ) ~ 126-135.
2.
Homotopy a s s o c i a t i v i t y of H-spaces I and 11. Trans. Amer. Math. Soc
3.
. 108 (1963) , 275-292
and 293-312.
Manifolds of t h e homotopy type of (non L i e ) groups.
Bull. h e r .
Math. SOC. 75 (1969), 998-1000.
4.
H-spaces and c l a s s i f y i n g spaces, foundations and r e c e n t developments.
5.
Proc
. Symp. Pure Math.
XXII (1970) , 247-272.
H-spaces from t h e homotopy point o f view. Mathematics.
161, Springer (1970).
Lecture n o t e s i n
217
Bibliography
6.
Sphere bundles over spheres as H-spaces
mod p > 2.
Symp. on
Algebraic Topology, B a t t e l l e S e a t t l e 1971, Lecture notes i n Math.
249, Springer, 106-110. and Epstein, D.B.A.,
Steenrod, N.E.
Studies 50.
Cohomology operations.
Princeton University Press
Ann. of Math.
1962.
Thomas, E. , 1.
On t h e
mod 2
cohomology of c e r t a i n H-spaces.
Comment. Math.
Helv. 37 (1962/63) , 132-140. Ann. o f Math.
77 (1963), 306-317.
2.
Steenrod squares and H-spaces.
3.
Steenrod squares and H-spaces 11. Ann. of Math. 81 (1965)~
473-495.
4.
Postnikov i n v a r i a n t s and higher order cohomology operations. Ann. o f Math. ( 2 ) 85 (19671, 184-217.
Whitehead, G.W.,
On mappings i n t o group l i k e spaces.
28 (1954)
Wilkerson, C .
Comment. Math. Helv.
320-328.
, Genus
and c a n c e l l a t i o n f o r H-spaces
theory and homotopy theory ( S e a t t l e 1974).
418, Springer.
, Localization
i n group
Lecture notes i n Math.
157-159.
Wilson, S . , The &spectrum for Brown-Peterson cohomology.
Comment. Math.
Helv. 48 (1973), 45-55. Zahler, R . ,
The Adams-Novikov s p e c t r a l sequence for t h e spheres.
Ann. of
Math. ( 2 ) , 96 (1972) , 480-504. Zabrodsky, A. 1.
,
Implications i n t h e cohomology o f H-spaces.
(1970 2.
9
363-375
-
Ill. J . of Math. 14
Secondary operations i n t h e module of indecomposables. Adv. Study I n s t . on Algebraic Topology.
Proc.
Aarhus Denmark (1970),
658-672. 3.
Homotopy a s s o c i a t i v i t y and f i n i t e CW complexes.
(1970) , 121-128.
Topology 9
Bibliography
218
4.
Cohomology operations and homotopy commutative H-spaces. Steenrod algebra and i t s applications.
The
Lecture notes i n
Mathematics 168, Springer (1970) , 308-317.
5.
On spherical classes i n t h e cohomology of H-spaces.
H-spaces
Neuchatel A k t 1970, Lecture notes i n Math. 196, Springer, 25-33.
6.
Secondary operations i n the cohomology of H-spaces.
I l l . J. o f
Math. 1 5 (1971), 648-655.
7.
The c l a s s i f i c a t i o n o f simply connected H-spaces with t h r e e c e l l s . 11, Math. Scand. 30 (1972), 211-222.
8.
On t h e construction of new f i n i t e CW H-spaces, Inventiones Math. 16 (19721, 260-266.
9.
p-equivalences and homotopy type l o c a l i z a t i o n i n group theory and homotopy theory ( S e a t t l e 1974).
Lecture notes i n Math. 418,
Springer. 10.
On t h e homotopy type o f p r i n i c p a l c l a s s i c a l group bundles over spheres.
Israel J . of Math. ll (1972), 315-325.
L i s t of symbols
219
The following a r e used as s t a n d a r d n o t a t i o n s throughout t h e s e n o t e s . We i n d i c a t e below t h e page where t h e y a r e f i r s t introduced.
A : X x X -> A : X ->
X X
x
X,
A
X,
t h e i d e n t i f i c a t i o n map
t h e diagonal
1
1
-
A = AA
-
re1 f
2, 18
core1 f
2, 18
[f]
homotopy class of
2
f
[XI homotopy type o f X [X,Y] = r [ f ] I f :
x
-+
Yl
h* = [ h , ~ ] g* = [X&l
( )#
t h e a d j o i n t of t h e homotopy pullback of
fo
t h e homotopy pushout of W
*,f
M
*,f
fo
and
and
fl
3
= fiber f
4
= C(f)
4
{En, Y 1, $ -spectrum l n 'n,m'
'En> 'n'
'n,m
1,
h r i n g &spectrum
F t h e f o l d i n g map X,y
fl
9
an H-space
9
L X t h e loop space of X E: X d
+
X,
t h e e v a l u a t i o n at
H f t h e category of H-spaces
5, 6
11, 12 t
11 15
L i s t of symbols
220
PX t h e Moore ( f r e e ) p a t h space
LX
t h e based Moore path space
CX t h e Moore loop space
a : PX
+ R+
t h e length of a path t h e e v a l u a t i o n at
Eo, Et, Em: PX + X kt: X + PX P f , Lf,
the length
t
0, t ,
constant p a t h
16
E
16
P X C P X x PX 2
Add: T2X + PX
16
t h e path a d d i t i o n
18
t h e p r e c i s e category of H-spaces
HfF
an H-map i n
f,F
15
18
H
fF
HA
a homotopy a s s o c i a t i v e H-space
1 9 , 58
HC
a homotopy commutative H-space
1 9 , 62
f+g =
!l*(f,d=
t h e d i f f e r e n c e between
Df,g
22
P ( f x g)A f
and
g
22 22
D = D p1'p2 t h e H-deviation of
HD(f,p,
p')
K(G,m)
Eilenberg-MacLane space
Htn
homotopy approximation
25
27 42
t h e Moore space
44
homology approximation
45
M( G,n)
Han
f
0 = B(f,F,A,A')
the HA o b s t r u c t i o n
59
KN
homotopy n i l p o t e n t
64
HS
homotopy solvable
64
PA, &A, Pi
p r i m i t i v e s and indecomposables
t h e Steenrod operations
71
,
L i s t of Symbols
a = (x,E,y)
a cohomology operation
P
t h e s e t of primes
M
Pl equivalence
( f . c ) , (f.c.T), f
X(P ,$), 1
r,
mod
P1
P r o p o s i t i o n 4.3.1
MiX(Xi'
h
t h e genus of
X
Pi, $i)
F(X) = F
(PI
f i n i t e n e s s condition
homotopic t o
r(1-1, 4
G(X)
98
114
(f.c.H)
f(Pl,$)
2 21
(X)
B(n,p) = F(p)(K(Zp,n))
g
115 126 128
222
Index of terminology
We i n d i c a t e below t h e page where t h e following notions a r e f i r s t introduced i n t h i s volume.
Action
47, 48
F i n i t e n e s s condition
115
Adams r e s o l u t i o n
190
Generic property
147
BP r e s o l u t i o n
193
Genus
1 47
Homology approximation
44
Adjoint maps
3
Bocks t e i n
Homotopy
operations
86
s p e c t r a l sequence
85
Browder's implication theorems 96, 97 Brown Peterson spectrum
163
Category of H-spaces
1 5 , 18
Cohomology
4 98
generalized operations operations of order Connective f i b e r i n g
k
98 43
Decomposition of 0-equivalence 128 Eckmann-Hi It on d u a l i t y
4
0-equivalence
1 1 4 , 115
Essentially different H-structures
27
Evaluation maps
15
category core1 f re1 f Homotopy approximation
42
Homotopy a s s o c i a t i v i t y
1 9 , 58
Homot opy commutativity
1 9 , 62
Homotopy nilpotency
64
Homot opy s o l v a b i l i t y
64
Hopf a l g e b r a
69, 70
H-action
48
H-deviation
35
H-homot opy equivalence
27
H - l i f t ings
37
H-mp
15
Index of terminology
H-space mod P 1
non c l a s s i c a l
Ho space
9 136 209 157 1239 124
Implications
96
Mixing homotopy types
152
Moore decomposition
44
Moore loop space
16
Moore path space
15
Moore space
44
Pos tnikov decomposition
43
p r i m i t i v e decomposition
44
system
42
Principal fibrations maps of
49 50
Pushout
3
Pullbacks
3
P1 equivalence
114
]El
homotopy
126
P1 universal
115
%spectrum
4
223
Ring spectrum
5
Weakly p r i n c i p a l (w-principal) Wilson's
B(n,p)
53 172
This Page Intentionally Left Blank