ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY Christian Ausoni and John Rognes Abstract. Let `p be the p-complete connective Adams summand of topologica* *l K- theory, with coefficient ring (`p)* = Zp[v1], and let V (1) be the Smith-* *Toda complex, with BP*(V (1)) = BP*=(p; v1). For p 5 we explicitly compute the V (1)-h* *omotopy of the algebraic K-theory spectrum of `p, denoted V (1)*K(`p). In partic* *ular we find that it is a free finitely generated module over the polynomial alge* *bra P (v2), except for a sporadic class in degree 2p - 3. Thus also in this case alg* *ebraic K- theory increases chromatic complexity by one. The proof uses the cyclotom* *ic trace map from algebraic K-theory to topological cyclic homology, and the calcu* *lation is actually made in the V (1)-homotopy of the topological cyclic homology of* * `p. Contents Introduction 1.Classes in algebraic K-theory 2.Topological Hochschild homology 3.Topological cyclotomy 4.Circle homotopy fixed points 5.The homotopy limit property 6.Higher fixed points 7.The restriction map 8.Topological cyclic homology 9.Algebraic K-theory Introduction We are interested in the arithmetic of ring spectra. To make sense of this we must work with structured ring spectra, such as S- algebras [EKMM], symmetric ring spectra [HSS] or -rings [Ly]. We will refer to these as S-algebras. The commutative objects are then commutative S-algebras. The category of rings is embedded in the category of S-algebras by the Eilen* *berg- Mac Lane functor R 7! HR. We may therefore view an S-algebra as a generalization of a ring in the algebraic sense. The added flexibility of S-algebras provides * *room for new examples and constructions, which may eventually also shed light upon t* *he category of rings itself. In algebraic number theory the arithmetic of the ring of integers in a numbe* *r field is largely captured by its Picard group, its unit group and its Brauer group. T* *hese are in turn reflected in the algebraic K-theory of the ring of integers. Algeb* *raic Typeset by AM S-T* *EX 1 2 CHRISTIAN AUSONI AND JOHN ROGNES K-theory is defined also in the generality of S-algebras. We can thus view the algebraic K-theory of an S-algebra as a carrier of some of its arithmetic prope* *rties. The algebraic K-theory of (connective) S-algebras can be closely approximated by diagrams built from the algebraic K-theory of rings [Du]. Hence we expect th* *at global structural properties enjoyed by algebraic K-theory as a functor of rings should also have an analogue for algebraic K-theory as a functor of S-algebras. We have in mind, in particular, theetale descent property of algebraic K-the* *ory conjectured by Lichtenbaum [Li] and Quillen [Qu], which has been established for several classes of commutative rings [V], [RW], [HM2]. We are thus led to ask w* *hen a map of commutative S-algebras A ! B should be considered as anetale covering with Galois group G. In such a situation we may further ask whether the natural map K(A) ! K(B)hG to the homotopy fixed point spectrum for G acting on K(B) induces an isomorphism on homotopy in sufficiently high degrees. These questions will be considered in more detail in [Ro2]. One aim of this line of inquiry is to find a conceptual description of the a* *lge- braic K-theory of the sphere spectrum, K(S0) = A(*), which coincides with Wald- hausen's algebraic K-theory of the one-point space *. In [Ro1] the second auth* *or computed the mod 2 spectrum cohomology of A(*) as a module over the Steenrod algebra, providing a very explicit description of this homotopy type. However, * *this result is achieved by indirect computation and comparison with topological cycl* *ic homology, rather than by a structural property of the algebraic K-theory functo* *r. What we are searching for here is a more memorable intrinsic explanation for the homotopy type appearing as the algebraic K-theory of an S-algebra. More generally, for a simplicial group G with classifying space X = BG there* * is an S-algebra S0[G], which can be thought of as a group ring over the sphere spe* *c- trum, and K(S0[G]) = A(X) is Waldhausen's algebraic K-theory of the space X. When X has the homotopy type of a manifold, A(X) carries information about the geometric topology of that manifold. Hence anetale descent description of K(S0[G]) will be of significant interest in geometric topology, reaching beyond* * al- gebraic K-theory itself. In the present paper we initiate a computational exploration of this `brave * *new world' of ring spectra and their arithmetic. We begin by considering some interesting examples of (pro-)etale coverings in the category of commutative S-algebras. For convenience we will choose to work locally, with S-algebras that are complete at a prime p. For the purpose of alg* *ebraic K-theory this is less of a restriction than it may seem at first. What we have* * in mind here is that the square diagram K(A) ________//K(Ap) | | | | fflffl| fflffl| K(ss0A) _____//_K(ss0Ap) is homotopy cartesian after p-adic completion [Du], when A is a connective S- algebra, Ap its p-completion, ss0A its ring of path components and ss0(Ap) ~=(s* *s0A)p. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 3 This reduces the p-adic comparison of K(A) and K(Ap) to the p-adic comparison of K(ss0A) and K(ss0Ap), i.e., to a question about ordinary rings, which we view as a simpler question, or at least as one lying in better explored territory. This leads us to study p-complete S-algebras, or algebras over the p-complete sphere spectrum S0p. This spectrum is approximated in the category of commutati* *ve S-algebras (or E1 ring spectra) by a tower of chromatic localizations [Ra1] S0p! . .!.LnS0p! . .!.L1S0p! L0S0p= HQp : Here Ln = LE(n) is Bousfield's localization functor [Bo], [EKMM] with respect to the nth Johnson-Wilson theory with coefficient ring E(n)* = Z(p)[v1; : :;:vn; v* *-1n]. By the Hopkins-Ravenel chromatic convergence theorem [Ra3, x8], the natural map S0p! holim nLnS0pis a homotopy equivalence. For each n 1 there is a further map of commutative S-algebras LnS0p! LK(n)S0pto the Bousfield localization with respect to the nth Morava K-theory with coefficient ring K(n)* = Fp[vn; v-1n]. * *This is an equivalence for n = 0, and LK(1)S0p' Jp is the non-connective p-complete image-of-J spectrum. There is a highly interesting sequence of commutative S-algebras En construc* *ted by Morava as spectra [Mo], by Hopkins and Miller [Re] as S-algebras (or A1 ring spectra) and by Goerss and Hopkins [GH] as commutative S-algebras (or E1 ring spectra). The coefficient ring of En is (En)* ~=W Fpn[[u1; : :;:un-1 ]][u; u-1 * *]. As a special case E1 ' KUp is the p-complete complex topological K-theory spectrum. The cited authors also construct a group action on En through commutative S- algebra maps, by a semidirect product Gn = Sn oCn where Sn is the nth (profinit* *e) Morava stabilizer group [Mo] and Cn = Gal(Fpn=Fp) is the cyclic group of order * *n. There is a homotopy equivalence LK(n)S0p' EhGnn, where the homotopy fixed point spectrum is formed in a continuous sense [DH], which reflects the Morava change of rings theorem [Mo]. Furthermore, the space of self equivalences of En in the category of commu- tative S-algebras is weakly equivalent to its group of path components, which is precisely Gn. In fact the extension LK(n)S0p! En qualifies as a pro-etale cover* *ing in the category of commutative S-algebras, with Galois group weakly equivalent * *to Gn. The weak contractibility of each path component of the space of self equiv- alences of En (over either S0por LK(n)S0p) serves as the commutative S-algebra version of the unique lifting property foretale coverings. Also the natural map i :En ! T HH(En) is a K(n)-equivalence, cf. [MS1, 5.1], implying that the space of relative K"ahler differentials of En over LK(n)S0pis contractible. See [Ro2* *] for further discussion. There are furtheretale coverings of En. For example there is one with coeff* *i- cient ring W Fpm [[u1; : :;:un-1 ]][u; u-1 ] for each multiple m of n. Let Enr* *nbe the colimit of these, with Enrn*= W Fp[[u1; : :;:un-1 ]][u; u-1 ]. Then Gal(Enrn=LK* *(n)S0p) is weakly equivalent to an extension of Sn by the profinite integers ^Z= Gal(Fp* * =Fp). Let En be a maximal pro-etale covering of En, and thus of LK(n)S0p. What is the absolute Galois group Gal(En =LK(n)S0p) of LK(n)S0p? The tower of commutative S-algebras induces a tower of algebraic K-theory sp* *ec- tra K(S0p) ! . .!.K(LnS0p) ! . .!.K(Jp) ! K(Qp) 4 CHRISTIAN AUSONI AND JOHN ROGNES studied in the p-local case by Waldhausen [Wa2]. The natural map K(S0p) ! holim nK(LnS0p) may well be an equivalence, see [MS2]. We are thus led to study the spectra K(LnS0p), and their relatives K(LK(n)S0p). (More precisely, Waldhau* *sen studied finite localization functors Lfncharacterized by their behavior on fini* *te CW- spectra. However, for n = 1 the localization functors L1 and Lf1agree, and this* * is the case that we will explore in the body of this paper. Hence we will suppress* * this distinction in the present discussion.) Granting that LK(n)S0p! En qualifies as anetale covering in the category of commutative S-algebras, the descent question concerns whether the natural map (0.1) K(LK(n)S0p) ! K(En)hGn is a p-adic homotopy equivalence in sufficiently high dimensions. We conjecture that it does. To analyze K(En) we expect to use a localization sequence in algebraic K-the* *ory to reduce to the algebraic K-theory of connective commutative S-algebras, and to use the B"okstedt-Hsiang-Madsen cyclotomic trace map to topological cyclic homology to compute these [BHM]. The ring spectra En and E(n)p are closely related, and for n 1 we expect that there is a cofiber sequence of spectra (0.2) K(BP p) ! K(BP p) ! K(E(n)p) analogous to the localization sequence K(Fp) ! K(Zp) ! K(Qp) in the case n = 0. Something similar should work for En. The cyclotomic trace map trc: K(BP p) ! T C(BP p; p) ' T C(BP ; p) induces a p-adic homotopy equivalence upon replacing the target with its connec* *tive cover [HM1]. Hence a calculation of T C(BP ; p) is as good as a calculation* * of K(BP p), after p-adic completion. In this paper we present computational tec* *h- niques which are well suited for calculating T C(BP ; p), at least when BP <* *n>p is a commutative S-algebra and the Smith-Toda complex V (n) exists as a ring spectrum. In the algebraic case n = 0, with BP <0> = HZ(p), these techniques si- multaneously provide a simplification of the argument in [BM1], [BM2] computing T C(Z; p) and K(Zp) for p 3. Presumably the simplification is related to that appearing in different generality in [HM2]. It is also plausible that variations on these techniques apply when replacin* *g V (n) by another finite type n + 1 ring spectrum, and the desired commutative S-algeb* *ra structure on BP p is weakened to the existence of an S-algebra map from a related commutative S-algebra, such as MU or BP . The first non-algebraic case occurs for n = 1. Then E1 ' KUp has an action by G1 = Zxp~= x . Here Zp ~= = 1 + pZp Zxp, Z=(p - 1) ~= Zxpand k 2 Zxp acts on E1 like the p-adic Adams operation k on KUp. Let Lp = Eh1 be the p-complete Adams summand with coefficient ring (Lp)* = Zp[v1; v-11], so Lp ' E(1)p. Then acts continuously on Lp with Jp ' Lhp . Let * *`p ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 5 be the p-complete connective Adams summand with coefficient ring (`p)* = Zp[v1], so `p ' BP <1>p. We expect that there is a cofiber sequence of spectra K(Zp) ! K(`p) ! K(Lp) : The previous calculation of T C(Z; p) [BM1], [BM2], and the calculation of T C(* *`; p) presented in this paper, identify the p-adic completions of K(Zp) and K(`p), re- spectively. Given an evaluation of the transfer map between them, this identif* *ies K(Lp). The homotopy fixed points for the -action on K(Lp) induced by the Adams operations k for k 2 1 + pZp should then model K(Jp) = K(LK(1)S0p). This brings us to the contents of the present paper. In x1 we produce two us* *eful classes K1 and K2 in the algebraic K-theory of `p. In x2 we compute the V (1)- homotopy of the topological Hochschild homology of `, simplifying the argument of [MS1]. In x3 we present notation concerning topological cyclic homology and the cyclotomic trace map of [BHM]. In x4 we make preparatory calculations in the spectrum homology of the S1-homotopy fixed points of T HH(`). These are applied in x5 to prove that the canonical map from the Cpn fixed points to the * *Cpn homotopy fixed points of T HH(`) induces an equivalence on V (1)-homotopy above dimension (2p - 2), using [Ts]. In x6 we inductively compute the V (1)-homotopy of all these (homotopy) fixed point spectra, and their homotopy limit T F (`; p* *). The action of the restriction map on this limit is then identified in x7. The p* *ieces of the calculation are brought together in Theorem 8.4 of x8, yielding an expli* *cit computation of the V (1)-homotopy of T C(`; p): Theorem 0.3. Let p 5. There is an isomorphism of E(1; 2) P (v2)-modules V (1)*T C(`; p)~=E(1; 2; @) P (v2) E(2) P (v2) Fp{1te | 0 < e < p} E(1) P (v2) Fp{2tep | 0 < e < p} with |1| = 2p - 1, |2| = 2p2 - 1, |v2| = 2p2 - 2, |@| = -1 and |t| = -2. The p-completed cyclotomic trace map K(`p)p ! T C(`p; p) ' T C(`; p) identif* *ies K(`p)p with the connective cover of T C(`; p). This yields the following expres* *sion for the V (1)-homotopy of K(`p), given in Theorem 9.1 of x9: Theorem 0.4. Let p 5. There is an exact sequence of E(1; 2)P (v2)-modules 0 ! 2p-3Fp -! V (1)*K(`p) -trc-!V (1)*T C(`; p) ! -1 Fp ! 0 taking the degree 2p - 3 generator in 2p-3HFp to a class a 2 V (1)2p-3K(`p), and taking the class @ in V (1)-1 T C(`; p) to the degree -1 generator in -1 HFp. The V (1)-homotopy of any spectrum is a P (v2)-module, but we emphasize that V (1)*T C(`; p) is a free finitely generated P (v2)-module, and V (1)*K(`p) is * *free and finitely generated except for the summand Fp{a} in degree 2p - 3. Hence both K(`p)p and T C(`; p) are fp-spectra in the sense of [MR], with finitely present* *ed 6 CHRISTIAN AUSONI AND JOHN ROGNES mod p cohomology as a module over the Steenrod algebra. They both have fp- type 2, because V (1)*K(`p) is infinite while V (2)*K(`p) is finite, and simila* *rly for T C(`; p). In particular, K(`p) is closely related to elliptic cohomology. More generally, at least if BP p is a commutative S-algebra and V (n) exi* *sts as a ring spectrum, similar calculations to those presented in this paper show * *that V (n)*T C(BP ; p) is a free P (vn+1 )-module on 2n+2 + 2n(n + 1)(p - 1) gene* *ra- tors. So algebraic K-theory takes such fp-type n commutative S-algebras to fp-t* *ype (n + 1) commutative S-algebras. If our ideas about localization sequences are c* *or- rect then also K(En)p will be of fp-type (n + 1), and ifetale descent holds in algebraic K-theory for LK(n)S0p! En with cdp(Gn) < 1 then also K(LK(n)S0p)p will be of fp-type (n + 1). The moral is that algebraic K-theory in many cases increments chromatic complexity by one, i.e., it produces a constant red-shift * *in stable homotopy theory. Notations and conventions. For an Fp vector space V , let E(V ), P (V ) and (V* * ) be the exterior algebra, polynomial algebra and divided power algebra on V , re* *spec- tively. When V has a basis {x1; : :;:xn}, we write E(x1; : :;:xn), P (x1; : :;:* *xn) and (x1; : :;:xn) for these algebras. So (x) = Fp{flj(x) | j 0} with fli(x) . flj(* *x) = (i; j)fli+j(x). Let Ph(x) = P (x)=(xh = 0) be the truncated polynomial algebra * *of height h. For a b 1 let Pab(x) = Fp{xk | a k b} as a P (x)-module. By an infinite cycle in a spectral sequence we mean a class x such that dr(x* *) = 0 for all r. By a permanent cycle we mean an infinite cycle which is not a bounda* *ry, i.e., a class that survives to represent a nonzero class at E1 . Differentials * *are often only given up to multiplication by a unit. 1. Classes in algebraic K-theory 1.1. E1 ring spectrum models. Let p be an odd prime. Let ` = BP <1> be the Adams summand of p-local connective topological K-theory. Its homotopy groups are `* ~=Z(p)[v1], with |v1| = q = 2p - 2. Its p-completion `p with `p* ~=Zp[v1] admits a model as an E1 ring spectrum, which can be constructed as the algebraic K-theory spectrum of a perfect field k0. Let g be a prime power topologically generating the p-adic units and let k0* * = colimn0 Fgpn k be a Zp-extension of k = Fg. Then `p = K(k0)p is an E1 ring spectrum model for the p-completed Adams summand. Likewise jp = K(k)p and kup = K(k)p are E1 ring spectrum models for the p-completed image-of-J spectrum and the p-completed connective topological K- theory spectrum, respectively. The Frobenius automorphism oeg(x) = xg induces the Adams operation g on both `p and kup. Then k is the fixed field of oeg, and jp is the connective cover of the homotopy fixed point spectrum for g acting on either one of `p or kup. The E1 ring spectrum maps S0p! jp ! `p ! kup ! HZp induce E1 ring spectrum maps on algebraic K-theory: K(S0p) -! K(jp) -! K(`p) -! K(kup) -! K(Zp) : In particular these are H1 ring spectrum maps. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 7 1.2. A first class in algebraic K-theory. The B"okstedt trace map tr :K(Zp) ! T HH(Zp) maps onto the first p-torsion in the image, which is T HH2p-1(Zp) ~= Z=p{e} [BM1]. Let eK 2 K2p-1(Zp) be a class with tr(eK ) = e. There is a (2p - 2)-connected linearization map `p ! HZp of E1 ring spectra, which induces a (2p - 1)-connected map K(`p) ! K(Zp). Definition 1.3. Let K1 2 K2p-1(`p) be a chosen class mapping to eK 2 K2p-1(Zp) under the map induced by linearization `p ! HZp. The image tr(K1) 2 T HH2p-1(`p) of this class under the trace map tr :K(`p) ! T HH(`p) will map under linearization to e 2 T HH2p-1(Zp). Remark 1.4. The class K1 2 K2p-1(`p) does not lift further back to K2p-1(S0p), since eK has a nonzero image in ss2p-2 of the homotopy fiber of K(S0p) ! K(Zp) [Wa1]. Thus K1 does also not lift to K2p-1(jp), because the map S0p! jp is (pq - 2)-connected. It is not clear if the induced action of g on K(`p) leaves* * K1 invariant. 1.5. Homotopy and homology operations. For a spectrum X, let DpX = Ep np X^p be its pth extended power. Part of the structure defining an H1 ring spectrum E is a map :DpE ! E. Then a mod p homotopy class 2 ssm (DpSn ; Fp) determines a mod p homotopy operation * :ssn(E) -! ssm (E; Fp) natural for maps of H1 ring spectra E. Its value on the homotopy class represe* *nted by a map a: Sn ! E is the image of under the composite map ssm (DpSn ; Fp) -Dp(a)---!ssm (DpE; Fp) -! ssm (E; Fp) : Likewise the Hurewicz image h() 2 Hm (DpSn ; Fp) induces a homology operation h()* :Hn(E; Fp) -! Hm (E; Fp) ; and the two operations are compatible under the Hurewicz homomorphisms. For Sn with n = 2k - 1 an odd dimensional sphere the bottom two cells of DpSn are in dimensions pn + (p - 2) and pn + (p - 1), and are connected by a mod p Bockstein. Hence the bottom two mod p homotopy classes of DpSn are in these two dimensions, and are called fiP kand P k, respectively. Their Hurewicz images induce the Dyer-Lashof operations denoted fiQk and Qk in homology, cf. [Br]. 1.6. A second class in algebraic K-theory. We use the H1 ring spectrum structure on K(`p) to produce a further element in its mod p homotopy. Definition 1.7. Let K2 = (P p)*(K1) 2 K2p2-1(`p; Fp) be the image under the mod p homotopy operation (P p)* :K2p-1(`p) -! K2p2-1(`p; Fp) of K1 2 K2p-1(`p) . 8 CHRISTIAN AUSONI AND JOHN ROGNES Since the trace map tr :K(`p) ! T HH(`p) is an E1 ring spectrum map, it fol- lows that tr(K2) 2 T HH2p2-1(`p; Fp) equals the image of tr(K1) 2 T HH2p-1(`p) under the mod p homotopy operation (P p)*. We shall identify this image in the next section. Remark 1.8. It is not clear whether K2 lifts to an integral homotopy class in K2p2-1(`p). The image of eK 2 K2p-1(Zp) in K2p-1(Qp; Fp) is v1d logp for a cla* *ss d logp 2 K1(Qp; Fp) mapping to the generator of K0(Fp; Fp) in the K-theory loca* *l- ization sequence for Zp, [HM2]. It appears that the image of K2 in V (1)2p2-1(L* *p) is v2d logv1 for a class d logv1 2 V (1)1K(Lp) mapping to the generator of V (0)0K* *(Zp) in the expected K-theory localization sequence for `p. The classes K1 and K2 are therefore related to logarithmic differentials for poles at p and v1, respectiv* *ely, which partially motivates the choice of the letter `'. 2. Topological Hochschild homology The topological Hochschild homology functor T HH(-), as well as its refined versions T HH(-)hS1, T HH(-)Cpn , T F (-; p) and T C(-; p), preserve p-adic equ* *iv- alences. Hence we will tend to write T HH(Z) and T HH(`) in place of T HH(Zp) and T HH(`p), and similarly for the other related functors. 2.1. Homology of T HH(`). The ring spectrum map ` ! HFp induces an injec- tion on mod p homology, identifying H*(`; Fp) with the subalgebra H*(`; Fp) = E(ok | k 2) P (k | k 1) of the dual Steenrod algebra A*. There is a B"okstedt spectral sequence (2.2) E2**= HH*(H*(`; Fp)) =) H*(T HH(`); Fp) with E2**= H*(`; Fp) (oeok | k 2) E(oek | k 1) : Here oex 2 HH1(-) is represented by 1 x in degree 1 of the Hochschild complex. It corresponds to the map oe :` ! T HH(`) induced by the S1 action on T HH(`) and the inclusion of 0-simplices ` ! T HH(`). By naturality with respect to the map ` ! HFp, the differentials dp-1 (flj(oeok)) = oek+1 . flj-p(oeok) for j p, found in the B"okstedt spectral sequence for T HH(Fp), lift to the sp* *ectral sequence (2.2) above. Hence Ep**= H*(`; Fp) Pp(oeok | k 2) E(oe1 ; oe2 ) and this equals the E1 -term for bidegree reasons. In H*(T HH(`); Fp) there are Dyer-Lashof operations acting, and (oeok)p = Qpk(oeok) = oe(Qpk(ok)) = oeok+1 for all k 2 [St]. Thus as an algebra (2.3) H*(T HH(`); Fp) ~=H*(`; Fp) E(1; 2) P () where 1, 2 and are represented by oe1 , oe2 and oeo2, respectively, in the B"okstedt spectral sequence. Here |1| = 2p - 1, |2| = 2p2 - 1 and || = 2p2. Furthermore Qp(1) = Qp(oe1 ) = oe(Qp(1 )) = oe2 , so we may assume that we have chosen 2 = Qp(1). ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 9 2.4. V (1)-homotopy of T HH(`). Let V (n) be the nth Smith-Toda complex, with homology H*(V (n); Fp) ~= E(o0; : :;:on). Thus V (-1) = S0, V (0) is the mod p Moore spectrum, and V (1) is the cofiber of the multiplication by v1-map qV (0) ! V (0). There are cofiber sequences S0 -p!S0 -i0!V (0) -j0!S1 and qV (0) -v1!V (0) -i1!V (1) -j1!q+1V (0) defining the maps labeled i0, j0, i1 and j1. When p 5, V (1) is a commutative ring spectrum [Ok]. For a spectrum X the rth (partially defined) v1-Bockstein homomorphism fi1;r is defined on the classes x 2 V (1)*(X) with j1(x) 2 V (0)*(X) divisible by vr-* *11. Then for y 2 V (0)*(X) with vr-11. y = j1(x) let fi1;r(x) = i1(y) 2 V (1)*(X). * * So fi1;rdecreases degrees by rq + 1. Definition 2.5. Let r(n) = 0 for n 0, and let r(n) = pn + r(n - 2) for all n * * 1. Thus r(2n - 1) = p2n-1 + . .+.p (n odd powers of p) and r(2n) = p2n + . .+.p2 (n even powers of p). Note that (p2 - 1)r(2n - 1) = p2n+1 - p, while (p2 - 1)r(2n)* * = p2n+2 - p2. Proposition 2.6 (McClure-Staffeldt). There is an algebra isomorphism V (1)*T HH(`) ~=E(1; 2) P () with |1| = 2p - 1, |2| = 2p2 - 1 and || = 2p2. There are v1-Bocksteins fi1;p() = 1, fi1;p2(p) = 2 and generally fi1;r(n)(pn-1) 6= 0 for n 1. Proof. One proof proceeds as follows, leaving the v1-Bockstein structure to the more detailed work of [MS1]. H*(T HH(`); Fp) is an A*-comodule algebra. The coaction :H*(T HH(`); Fp) ! A* H*(T HH(`); Fp) agrees with the coproduct :A* ! A* A* when both are restricted to the subalgebra H*(`; Fp) A*. Furthermore (oex) = (1 oe) (x), so (1) = 1 1, (2) = 1 2 and () = 1 + o0 2. There are change-of-rings isomorphisms of A*-comodule algebras H*(V (1) ^ T HH(`); Fp) ~=(A*H*(`;Fp)Fp) (H*(`; Fp) E(1; 2) P ()) ~=A*H*(`;Fp)(H*(`; Fp) E(1; 2) P ()) ~=A* E(1; 2) P () : The homology classes 1 ^ 1, 1 ^ 2 and 1 ^ - o0^ 2 in H*(V (1) ^ T HH(`); Fp) are primitive, and correspond to the classes 1, 2 and under the isomorphisms above. Hence these three homology classes are in the Hurewicz image from spheri* *cal classes 1 2 V (1)2p-1T HH(`), 2 2 V (1)2p2-1T HH(`) and 2 V (1)2p2T HH(`), respectively. 10 CHRISTIAN AUSONI AND JOHN ROGNES Proposition 2.7. The classes K1 2 K2p-1(`p) and K2 2 K2p2-1(`p; Fp) map under the trace map to integral and mod p lifts of 1 2 V (1)2p-1T HH(`) and 2 2 V (1)2p2-1T HH(`), respectively. Proof. The linearized image in V (1)2p-1T HH(Z) of 1 2 V (1)2p-1T HH(`) equals the mod p and v1 reduction i1i0(e) of the class e 2 T HH2p-1(Z), since both cla* *sses have the same Hurewicz image 1 ^ oe1 in H2p-1(V (1) ^ T HH(Z); Fp). Thus the mod p and v1 reduction of the trace image tr(K1) equals 1. The Hurewicz image in H2p2-1(T HH(`); Fp) of tr(K2) = (P p)*(tr(K1)) equals the image of the homology operation Qp on the Hurewicz image 1 = oe1 in H2p-1(T HH(`); Fp) of tr(K1). This is Qp(oe1 ) = oeQp(1 ) = oe2 = 2. So the mod v1 image in V (1)2p2-1T HH(`) of tr(K2) equals 2, since both classes have the same Hurewicz image 1 ^ 2 in H2p2-1(V (1) ^ T HH(`); Fp). 3. Topological cyclotomy We now review some terminology and notation concerning topological cyclic homology and the cyclotomic trace map. 3.1. Frobenius, restriction, Verschiebung. As already indicated, T HH(`) is an S1-equivariant spectrum. Let Cpn S1 be the cyclic group of order pn. The Frobenius maps F :T HH(`)Cpn ! T HH(`)Cpn-1 are the usual inclusions of fixed point spectra that forget part of the invariance. Their homotopy limit defines T F (`; p) = holimn;FT HH(`)Cpn : There are also restriction maps R :T HH(`)Cpn ! T HH(`)Cpn-1 , defined using the cyclotomic structure of T HH(`), cf. [HM1]. They commute with the Frobenius maps, and thus induce a self map R :T F (`; p) ! T F (`; p). Its homotopy equa* *l- izer with the identity map defines the topological cyclic homology of `, which * *was introduced in [BHM]: ___R__// T C(`; p)__ss_//_T F (`;_p)__//T F (`; p) : 1 Hence there is a cofiber sequence -1 T F (`; p) -@!T C(`; p) -ss!T F (`; p) -1-R--!T F (`; p) ; which we shall use to compute V (1)*T C(`; p). There are also Verschiebung maps V :T HH(`)Cpn-1 ! T HH(`)Cpn , defined up to homotopy in terms of the S1- equivariant transfer. 3.2. The cyclotomic trace map. The B"okstedt trace map admits lifts trn :K(`p) ! T HH(`)Cpn for all n 0, with tr = tr0, which commute with the Frobenius maps and homotopy commute with the restriction maps up to preferred homotopy. Hence the limiting map trF :K(`p) ! T F (`; p) homotopy equalizes R and the identity map, and the resulting lift trc: K(`p) ! T C(`; p) is the B"okstedt-Hsiang-Madsen cyclotomic trace map [BHM]. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 11 3.3. The norm-restriction sequences. For each n 1 there is a homotopy commutative diagram K(`p) R RRRR trn|| RRtrn-1RRRRR N fflffl| R RR(( (3.4) T HH(`)hCpn _____//_T HH(`)Cpn________//T HH(`)Cpn-1 || | ^| || |n n| || h fflffl| h fflffl| T HH(`)hCpn _N___//_T HH(`)hCpn_R___//_^H(Cpn; T HH(`)) : The lower part is the map of cofiber sequences that arises by smashing the S1- equivariant cofiber sequence ES1+ ! S0 ! EeS1 with the S1-equivariant map T HH(`) ! F (ES1+; T HH(`)) and taking Cpn fixed point spectra. For closed subgroups G S1 recall that T HH(`)hG = F (ES1+; T HH(`))G is the G homotopy fixed point spectrum of T HH(`), and ^H(G; T HH(`)) = [EeS1 ^ F (ES1+; T HH(`)]G is the G Tate construction on T HH(`). The remaining terms of the diagram are then identified by the homotopy equivalences T HH(`)hCpn ' [ES1+^ T HH(`)]Cpn ' [ES1+^ F (ES1+; T HH(`))]Cpn and T HH(`)Cpn-1 ' [EeS1 ^ T HH(`)]Cpn : We call N, R, Nh and Rh the norm, restriction, homotopy norm and homotopy re- striction maps, respectively. We call n and ^n the canonical maps. The middle a* *nd lower cofiber sequences are the norm-restriction and homotopy norm-restriction * *se- quences, respectively. By passage to homotopy limits over Frobenius maps, there is also a limiting diagram K(`p) Q QQQQ trF || QtrFQQQQQ N fflffl| R QQ(( (3.5) T HH(`)hS1 _______//_T F (`;_p)________//T F (`; p) || | | || | |^ || Nh fflffl|1Rh fflffl| T HH(`)hS1 _____//_T HH(`)hS _____//_^H(S1; T HH(`)) : Implicit here are the p-adic homotopy equivalences T HH(`)hS1 ' holimn;FT HH(`)hCpn 1 hC n T HH(`)hS ' holimn;FT HH(`) p ^H(S1; T HH(`))' holim ^H(Cpn; T HH(`)) : n;F 12 CHRISTIAN AUSONI AND JOHN ROGNES 4. Circle homotopy fixed points 4.1. The circle trace map. The circle trace map 1 1 S1 trS1 = O trF :K(`p) ! T HH(`)hS = F (ES+ ; T HH(`)) is a preferred lift of the trace map tr :K(`p) ! T HH(`). We take S1 as our mo* *del for ES1. Let 1 T n= F (S1 =S2n-1 ; T HH(`))S for n 0, so that there is a descending filtration {T n}n on T 0= T HH(`)hS1, w* *ith layers T n=T n+1 ~=F (S2n+1 =S2n-1 ; T HH(`))S1 ~=-2n T HH(`). 4.2. The homology spectral sequence. Placing T n in filtration s = -2n and applying homology, we obtain a (not necessarily convergent) homology spectral sequence 1 (4.3) E2s;t= H-s (S1; Ht(T HH(`); Fp)) =) Hs+t(T HH(`)hS ; Fp) with E2**= P (t) H*(`; Fp) E(1; 2) P () : Here t has bidegree (-2; 0) while the other generators are located on the verti- cal axis. (No confusion should arise from the double usage of t as a polynomial cohomology class and the vertical degree in this or other spectral sequences.) Proposition 4.4. There are differentials d2(1 ) = t1, d2(2 ) = t2, d2(o2) = t and d2p(p1) = tp2 in the spectral sequence (4.3). Proof. The d2-differential d20;t:E20;t~=Ht(T HH(`); Fp){1} -! E2-2;t+1~=Ht+1(T HH(`); Fp){t} is adjoint to the S1-action on T HH(`), hence restricts to oe on Ht(`; Fp). Th* *us d2(i ) = toei = ti for i = 1 and 2, and d2(o2) = toeo2 = t. Write q = 2p-2 and let x 2 Cq(T 0; T 1; Fp) be a chain representing the diff* *erential d2(1 ) = t1, i.e., x maps to a cycle representing 1 in Hq(T HH(`); Fp) = E20;2* *p-2, and has boundary dx 2 Cq-1 (T 1; Fp) mapping to a cycle representing t1 in Hq-1 (-2 T HH(`); Fp) = E2-2;2p-1. Let :DpT 0 ! T 0 be the pth H1 struc- ture map for T 0= T HH(`)hS1, and form the chain *(e0 xp ) 2 Cpq(T 0; T p; Fp* *). It maps to a cycle representing p1in Hpq(T HH(`); Fp) = E20;2p2-2p, and has bou* *nd- ary *(e0 d(xp )) 2 Cpq-1(T p; Fp). By a chain level calculation [Br, 3.4] in * *the extended pth power of the pair (D2p-2; S2p-3 ), this boundary is homologous to a unit multiple of *(ep-1 (dx)p ), representing Qp-1 (t1) = tpQp(1) = tp2 in Hpq-1(-2p T HH(`); Fp) = E2-2p;2p2-1. Hence there is a nontrivial different* *ial d2p(p1) = tp2 in the spectral sequence above. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 13 4.5. The V (1)-homotopy spectral sequence. Applying V (1)-homotopy to the filtration {T n}n, in place of homology, we obtain a conditionally convergent V* * (1)- homotopy spectral sequence 1 (4.6) E2s;t(S1) = H-s (S1; V (1)tT HH(`)) =) V (1)s+tT HH(`)hS with E2**(S1) = P (t) E(1; 2) P () : Again t has bidegree (-2; 0) while the other generators are located on the vert* *ical axis. Definition 4.7. Let ff1 2 ss2p-3(S0), fi012 ss2p2-2p-1V (0) and v2 2 ss2p2-2V (1) be the classes represented in their respective Adams spectral sequences by the cobar cycles h10 = [1 ], h11 = [p1] and [o2]. So j1(v2) = fi01and j0(fi01) = f* *i1 2 ss2p2-2p-2(S0). Consider the unit map S0 ! K(`p) ! T HH(`)hS1, which is well defined after p-adic completion. Proposition 4.8. The classes i1i0(ff1) 2 ss2p-3V (1), i1(fi01) 2 ss2p2-2p-1V (* *1) and v2 2 ss2p2-2V (1) map under the unit map V (1)*S0 ! V (1)*T HH(`)hS1 to classes represented in E1 (S1) by t1, tp2 and t, respectively. Proof. Consider first the filtration subquotient T 0=T 2= F (S3+; T HH(`))S1. * *The unit map V (1) ! V (1) ^ (T 0=T 2) induces a map of Adams spectral sequences, taking the permanent cycles [1 ] and [o2] in the source Adams spectral sequence to infinite cycles with the same cobar names in the target Adams spectral se- quence. These are not boundaries in the cobar complex for the A*-comodule H*(T 0=T 2; Fp), because of the differentials d2(1 ) = t1 and d2(o2) = t that are present in the two-column spectral sequence converging to H*(T 0=T 2; Fp). * *In detail, H2p-2(T 0=T 2; Fp) = 0 and H2p2-1(T 0=T 2; Fp) is spanned by the primit* *ives 2 and 1p1. Thus [1 ] and [o2] are nonzero infinite cycles in the target Adams E2-term. * *They have Adams filtration one, hence cannot be boundaries. Thus they are permanent cycles, and the classes i1i0(ff1) and v2 have nonzero images under the composite V (1)* ! V (1)*(T 0) ! V (1)*(T 0=T 2). Thus they are also detected in V (1)*(T* * 0), in filtration s -2. For bidegree reasons the only possibility is that i1i0(ff* *1) is detected in the V (1)-homotopy spectral sequence E1 (S1) as t1, and v2 is detec* *ted as t. 1 Next consider the filtration subquotient T 0=T p+1 = F (S2p+1+; T HH(`))S . * *The unit map V (1) ! V (1) ^ (T 0=T p+1) again induces a map of Adams spectral se- quences, taking the permanent cycle [p1] in the source Adams spectral sequence to an infinite cycle with the same name. Again this is not a boundary in the cobar complex for the A*-comodule H*(T 0=T p+1; Fp), because of the differential 14 CHRISTIAN AUSONI AND JOHN ROGNES d2p(p1) = tp2 that is present in the (p + 1) column spectral sequence converging to H*(T 0=T p+1; Fp). Thus [p1] is a nonzero infinite cycle in the target Adams E2-term, of Adams filtration one. Hence the class i1(fi01) has a nonzero image under the composi* *te V (1)* ! V (1)*(T 0) ! V (1)*(T 0=T p+1). Thus it is also detected in V (1)*(T * *0) in filtration s -2p. Again for bidegree reasons the only possibility is that i1(f* *i01) is detected in the V (1)-homotopy spectral sequence E1 (S1) as tp2. 5. The homotopy limit property 5.1. Homotopy fixed point and Tate spectral sequences. For closed sub- groups G S1 we will consider the (second quadrant) G homotopy fixed point spectral sequence E2s;t(G)= H-s (G; V (1)tT HH(`)) =) V (1)s+tT HH(`)hG : We also consider the (upper half plane) G Tate spectral sequence ^E2s;t(G)= ^H-s(G; V (1)tT HH(`)) =) V (1)s+tH^(G; T HH(`)) : When G = S1 we have E2**(S1) = E(1; 2) P (t; ) since H*(S1; Fp) = P (t), and ^E2**(S1) = E(1; 2) P (t; t-1 ; ) since H^*(S1; Fp) = P (t; t-1 ). When G = Cpn we have E2**(Cpn) = E(un; 1; 2) P (t; ) since H*(Cpn; Fp) = E(un) P (t), while E^2**(Cpn) = E(un; 1; 2) P (t; t-1 ; ) since H^*(Cpn; Fp) = E(un) P (t; t-1 ). The homotopy restriction map Rh induces a map of spectral sequences E*(Rh): E*(G) ! ^E*(G) ; which on E2-terms identifies E2(G) with the restriction of E^2(G) to the second quadrant. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 15 The Frobenius and Verschiebung maps F and V are compatible under n and n-1 with homotopy Frobenius and Verschiebung maps F h and V h that induce maps of homotopy fixed point spectral sequences E*(F h): E*(Cpn) ! E*(Cpn-1) and E*(V h): E*(Cpn-1) ! E*(Cpn) : Here E2(F h) maps the even columns isomorphically and the odd columns trivially. On the other hand, E2(V h) maps the odd columns isomorphically and the even columns trivially. This pattern persists to higher Er-terms, until a different* *ial of odd length appears in either spectral sequence. Thus the spectral sequences E*(Cpn) and E*(Cpn-1) are abstractly isomorphic up to and including the Er- term, where r is the length of the first odd differential in either spectral se* *quence. The same remarks apply for the Tate spectral sequences. 5.2. Input for Tsalidis' theorem. Definition 5.3. A map A* ! B* of graded groups is k-coconnected if it is an isomorphism in all dimensions > k and injective in dimension k. Theorem 5.4. The canonical map ^1: T HH(`) ! ^H(Cp; T HH(`)) induces a (2p - 2)-coconnected map on V (1)-homotopy. Proof. Consider diagram (3:4) in the case n = 1. The classes i1i0(ff1), i1(fi01* *) and v2 in V (1)* map through V (1)*K(`p) and 1 O tr1 to classes in V (1)*T HH(`)hCp that are detected by t1, tp2 and t in E1 (Cp), respectively. Continuing by Rh to V (1)*H^(Cp; T HH(`)) these classes factor through V (1)*T HH(`), where they pass through zero groups. Hence the images of t1, tp2 and t in E^1 (Cp) must be zero, i.e., these infinite cycles in E^2(Cp) are boundaries. For dimension r* *easons the only possibilities are d2p(t1-p )= t1 2 p-p2 p d2p (t )= t 2 2+1 -p2 d2p (u1t )= t : The classes i1i0(K1) and i1(K2) in V (1)*K(`p) map by 1 O tr1 to classes in V (1)*T HH(`)hCp that have Frobenius images 1 and 2 in V (1)*T HH(`), and hence survive as permanent cycles in E10;*(Cp). Thus their images 1 and 2 in E^*(Cp) are infinite cycles. Hence the various Er-terms of the Cp Tate spectral sequence are: ^E2(Cp)= E(u1; 1; 2) P (t; t-1 ; t) E^2p+1(Cp) = E(u1; 1; 2) P (tp; t-p ; t) E^2p2+1(Cp) = E(u1; 1; 2) P (tp2; t-p2 ; t) E^2p2+2(Cp) = E(1; 2) P (tp2; t-p2 ) : 16 CHRISTIAN AUSONI AND JOHN ROGNES For bidegree reasons there are no further differentials, so E^2p2+2(Cp) = E^1 (* *Cp) and the classes 1, 2 and tp2 are permanent cycles. The map ^1 :T HH(`) ! ^H(Cp; T HH(`)) induces on V (1)-homotopy the homo- morphism 2 2 E(1; 2) P () -! E(1; 2) P (tp ; t-p ) that maps 1 7! 1, 2 7! 2 and 7! t-p2 . For the classes i1i0(K1) and i1(K2) in V (1)*K(`p) map by tr to 1 and 2 in V (1)*T HH(`), and by Rh O 1 O tr1 to the classes in V (1)*H^(Cp; T HH(`)) represented by 1 and 2. The class in V (1)*T HH(`) must have nonzero image in V (1)*H^(Cp; T HH(`)), since its pth v1-Bockstein fi1;p() = 1 has nonzero image there. Thus maps to the class represented by t-p2 , up to a unit multiple which we ignore. So V (1)*^1 is an isomorphism in dimensions greater than |12tp2| = 2p - 2, and is injective in dimension 2p - 2. 5.5. The homotopy limit property. Theorem 5.6. The canonical maps n : T HH(`)Cpn ! T HH(`)hCpn ^n : T HH(`)Cpn-1 ! ^H(Cpn; T HH(`)) and 1 : T F (`; p) ! T HH(`)hS ^: T F (`; p) ! ^H(S1; T HH(`)) all induce (2p - 2)-coconnected maps on V (1)-homotopy. Proof. The claims for n and ^n follows from 5.4 and a theorem of Tsalidis [Ts]. The claims for and ^ follow by1passage to homotopy limits, using the p-adic homotopy equivalence T HH(`)hS ' holim n;FT HH(`)hCpn and its analogue for the Tate constructions. 6. Higher fixed points Let [k] = 1 when k is odd, and [k] = 2 when k is even. Let 0[k]= [k+1], so t* *hat {[k]; 0[k]} = {1; 2} for all k. We write vp(k) for the p-valuation of k, i.e.,* * the exponent of the greatest power of p that divides k. By convention, vp(0) = +1. Recall the integers r(n) from 2.5. Theorem 6.1. In the Cpn Tate spectral sequence E^*(Cpn) there are differentials k-1-pk pk-1 r(k-2) d2r(k)(tp ) = [k]t (t) for all 1 k 2n, and 2n r(2n-2)+1 d2r(2n)+1(unt-p ) = (t) : ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 17 The classes 1, 2 and t are infinite cycles. We shall prove this by induction on n, the case n = 1 being settled in the p* *revious section. Hence we assume the theorem holds for one n 1 and we will establish i* *ts assertions for n + 1. The terms of the Tate spectral sequence are E^2r(m)+1(Cpn) = E(un; 1; 2) P (tpm ; t-pm ; t) Mm E(un; 0[k]) Pr(k-2)(t) Fp{[k]ti | vp(i) = k - 1} k=3 for 1 m 2n. Next ^E2r(2n)+2(Cpn)= E(1; 2) Pr(2n-2)+1(t) P (tp2n; t-p2n ) 2nM E(un; 0[k]) Pr(k-2)(t) Fp{[k]ti | vp(i) = k - 1} : k=3 For bidegree reasons the remaining differentials are zero, so E^2r(2n)+2(Cpn) = E^1 (Cpn), and the classes tp2n are permanent cycles. Proposition 6.2. The associated graded of V (1)*H^(Cpn; T HH(`)) is E^1 (Cpn) = E(1; 2) Pr(2n-2)+1(t) P (tp2n; t-p2n ) 2nM E(un; 0[k]) Pr(k-2)(t) Fp{[k]ti | vp(i) = k - 1} : k=3 Comparing E*(Cpn) with E^*(Cpn) via the homotopy restriction map Rh, we obtain the following: Proposition 6.3. In the Cpn homotopy fixed point spectral sequence E*(Cpn) the* *re are differentials k-1 pk+pk-1 r(k-2) d2r(k)(tp ) = [k]t (t) for all 1 k 2n, and 2n r(2n-2)+1 d2r(2n)+1(un) = tp (t) : The classes 1, 2 and t are infinite cycles. We also consider the -inverted spectral sequences -1 E*(G) for G closed in S* *1, obtained by inverting t in E^*(G) and restricting to the left half plane. The E* *2- term -1 E2(G) is obtained from E2(G) by inverting . At each term, the natural map E*(G) ! -1 E*(G) is an isomorphism in total degrees greater than 2p - 2, and an injection in total degree 2p - 2. 18 CHRISTIAN AUSONI AND JOHN ROGNES Proposition 6.4. In the -inverted spectral sequence -1 E*(Cpn) there are dif- ferentials k-pk-1 r(k)-pk-1 d2r(k)(p ) = [k](t) for all 1 k 2n, and 2n r(2n)+1 d2r(2n)+1(unp ) = (t) : The classes 1, 2 and t are infinite cycles. The terms of the -inverted spectral sequence are m -pm -1 E2r(m)+1(Cpn) = E(un; 1; 2) P (p ; ; t) Mm E(un; 0[k]) Pr(k)(t) Fp{[k]j | vp(j) = k - 1} k=1 for 1 m 2n. Next 2n -p2n -1 E2r(2n)+2(Cpn) = E(1; 2) Pr(2n)+1(t) P (p ; ) 2nM E(un; 0[k]) Pr(k)(t) Fp{[k]j | vp(j) = k - 1} : k=1 Again -1 E2r(2n)+2(Cpn) = -1 E1 (Cpn) for bidegree reasons, and the classes p2n are permanent cycles. Proposition 6.5. The associated graded E1 (Cpn) of V (1)*T HH(`)hCpn maps by a (2p - 2)-coconnected map to 2n -p2n -1 E1 (Cpn) = E(1; 2) Pr(2n)+1(t) P (p ; ) 2nM E(un; 0[k]) Pr(k)(t) Fp{[k]j | vp(j) = k - 1} : k=1 Proof of Theorem 6.1. By our inductive hypothesis, the abutment -1 E1 (Cpn) contains summands 2n-2 p2n-1 p2n Pr(2n-1)(t){1p }, Pr(2n)(t){2 } and Pr(2n)+1(t){ } representing elements in V (1)*T HH(`)Cpn . By inspection there are no perma- nent cycles in the same total degree and of lower s-filtration in -1 E1 (Cpn) t* *han (t)r(2n-1). 1p2n-2, (t)r(2n). 2p2n-1 and (t)r(2n)+1. p2n, respectively. So the three homotopy classes represented by 1p2n-2, 2p2n-1 and p2n are v2-torsion classes of orders precisely r(2n - 1), r(2n) and r(2n) + 1, respectively. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 19 Consider the commutative diagram n ^n+1 T HH(`)hCpn oo____T HH(`)Cpn _____//_^H(Cpn+1; T HH(`)) |Fn| |Fn| |Fn| |fflffl 0 fflffl| ^1 fflffl| T HH(`) oo__=_____T HH(`) ________//_^H(Cp; T HH(`)) : Here F n is the n-fold Frobenius map forgetting Cpn-invariance. The right hand diagram commutes because ^n+1 is constructed as the Cpn-invariant part of an S1-equivariant model for ^1. The above three homotopy classes in V (1)*T HH(`)Cpn map by the middle F n to homotopy classes in V (1)*T HH(`) with the same names, and by ^1 to homotopy classes in V (1)*H^(Cp; T HH(`)) represented by 1t-p2n , 2t-p2n+1 and t-p2n+2 in E^1 (Cp), respectively. Hence they map by ^n+1 to permanent cycles in ^E*(Cpn+* *1) with these images under the right hand F n. By comparison over homotopy Frobenius and Verschiebung maps, there are iso- morphisms E^r(Cpn) ~=E^r(Cpn+1) for all r 2r(2n) + 1, taking un to un+1 . This determines the dr-differentials and Er-terms of ^E*(Cpn+1) up to and including * *the Er-term with r = 2r(2n) + 1: E^2r(2n)+1(Cpn+1) = E(un+1 ; 1; 2) P (tp2n; t-p2n ; t) 2nM E(un+1 ; 0[k]) Pr(k-2)(t) Fp{[k]ti | vp(i) = k - 1} : k=3 By inspection there are no permanent2cyclesnin2thensame+total1degree2andnof+hig* *her2 s-filtration in E^*(Cpn+1) than 1t-p , 2t-p and t-p , respectively. So also the equivalence ^n+1 -1ntakes the homotopy classes represented by 1p2n-2, 2p2n-12andnp2n+to2homotopy classes represented by 1t-p2n , 2t-p2n+1 and t-p , respectively. Since ^n+1 -1ninduces an isomorphism on V (1)-homotopy in dimensions > (2p- 2), it preserves the v2-torsion order of these classes. Thus the infinite cycles 2n r(2n) -p2n+1 r(2n)+1 -p2n+2 (t)r(2n-1). 1t-p , (t) . 2t and (t) . t are all boundaries in ^E*(Cpn+1). These are all t-periodic classes in ^Er(Cpn+1* *) for r = 2r(2n) + 1, hence cannot be hit by differentials on the t-torsion classes i* *n this Er-term. 2n 2n This leaves the t-periodic part E(un+1 ; 1; 2) P (tp ; t-p ; t), where a* *ll the generators above the horizontal2axisnare2infinitencycles.+Hence1the2differe* *ntialsn+2 hitting (t)r(2n-1) . 1t-p , (t)r(2n). 2t-p and (t)r(2n)+1 . t-p must originate on the horizontal axis, and the only possibilities are 2n-p2n+1 r(2n-1) -p2n d2r(2n+1)(t-p )= (t) . 1t 2n+1-p2n+2 r(2n) -p2n+1 d2r(2n+2)(t-p )= (t) . 2t 2n+2 r(2n)+1 -p2n+2 d2r(2n+2)+1(un+1 t-2p )= (t) . t : 20 CHRISTIAN AUSONI AND JOHN ROGNES The algebra structure on E^*(Cpn+1) allows us to rewrite these differentials as* * the remaining differentials asserted by case n + 1 of 6.1. Passing to the limit over the Frobenius maps, we obtain: Theorem 6.6. The associated graded of V (1)*H^(S1; T HH(`)) is ^E1(S1) = E(1; 2) P (t) M E(0[k]) Pr(k-2)(t) Fp{[k]ti | vp(i) = k - 1} : k3 Theorem 6.7. The associated graded E1 (S1) of V (1)*T HH(`)hS1 maps by a (2p - 2)-coconnected map to -1 E1 (S1) = E(1; 2) P (t) M E(0[k]) Pr(k)(t) Fp{[k]j | vp(j) = k - 1} : k1 Each of these E1 terms compute V (1)*T F (`; p) in dimensions > (2p - 2), by way of the (2p - 2)-coconnected maps and ^, respectively. 7. The restriction map We now evaluate the homomorphism R* :V (1)*T F (`; p) ! V (1)*T F (`; p) induced on V (1)-homotopy by the restriction map R, in dimensions > (2p - 2). The source and target are both identified with V (1)*T HH(`)hS1 via *. Then R* is identified with the composite homomorphism (^-1 )* O Rh*, where Rh is the homotopy restriction map. The latter induces a map of spectral sequences E*(Rh): E*(S1) ! ^E*(S1) ; where the E1 terms are given in 6.6 and 6.7. Proposition 7.1. In dimensions > (2p - 2) the homomorphism E1 (Rh) maps: (a) E(1; 2)P (t) in E1 (S1) isomorphically to E(1; 2)P (t) in ^E1 (S1). (b) E(0[k])Pr(k)(t)Fp{[k]-epk-1} in E1 (S1) onto E(0[k])Pr(k-2)(t) Fp{[k]tepk-1} in E^1 (S1), for k 3 and 0 < e < p. (c) the remaining terms in E1 (S1) to zero. Proof. Case (a) is clear. For (b) and (c) note that E1 (Rh) maps the term k-1 E(0[k]) Pr(k)(t) Fp{[k]-ep } in E1 (S1) to the term k-1 E(0[k]) Pr(k-2)(t) Fp{[k]tep } in E^1 (S1). Here e is prime to p. For e > p the source and target are in negat* *ive dimensions, while for e < 0 the source and target are concentrated in disjoint dimensions. The cases 0 < e < p remain, when the map is a surjection since r(k) - epk-1 > r(k - 2). This identifies the image of Rh*, by the following lemma from [BM1, x2]. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 21 Lemma 7.2. The representatives in E1 (S1) of the kernel of Rh*equal the kernel of E1 (Rh). Hence the image of Rh*is isomorphic to the image of E1 (Rh). The composite equivalence ^-1 does not induce a map of spectral sequences. Nonetheless it induces an isomorphism of E(1; 2) P (v2)-modules on V (1)- homotopy in dimensions > (2p - 2). Here v2 acts by multiplication in V (1)*, while multiplications by 1 and 2 are realized by the images of K1 and K2, since both and ^ are ring spectrum maps. Proposition 7.3. In dimensions > (2p - 2) the composite equivalence ^-1 in- duces an isomorphism 1 V (1)*H^(S1; T HH(`)) ~=V (1)*T HH(`)hS of P (v2)-modules, taking all classes represented by ffl11ffl22(t)m ti in E^1 (* *S1) to classes represented by ffl11ffl22(t)m j in E1 (S1) with i+p2j = 0. Here 0 ffl1* *; ffl2 1 and m 0. Proof. Let T^ = ^H(Cp; T HH(`)). The S1-equivariant map ^1 :T HH(`) ! T^ in- duces a map of S1 homotopy fixed points, which realizes the localization homo- morphism E*(S1) ! -1 E*(S1) on the level of spectral sequences. It follows that there is a homotopy equivalence V (1)^H^(S1; T HH(`)) ' V (1)^T^hS1 which agrees with ^-1 on (2p - 2)-connected covers. The v2-indivisible elements in V (1)*H^(S1; T HH(`)) are represented in E^1 * *(S1) by the classes E^1 (S1)=(t) = E(1; 2) M E(0[k]) Fp{[k]ti | vp(i) = k - 1} : k3 The v2-indivisible elements in V (1)*T^hS1 are represented in -1 E1 (S1) by the classes -1 E1 (S1)=(t) = E(1; 2) M E(0[k]) Fp{[k]j | vp(j) = k - 1} : k1 The asserted homotopy equivalence induces an isomorphism between these two terms, which by a dimension count must be given by ffl11ffl22ti 7- ! ffl11ffl22j with i + p2j = 0. Hence the same formulas hold modulo multiples of v2 on V (1)- homotopy. Taking the P (v2)-module structure into account, the corresponding formulas including factors (t)m also hold, and express the isomorphism 1 V (1)*H^(S1; T HH(`)) ~=V (1)*T^hS which agrees with ^-1 in dimensions > (2p - 2). 22 CHRISTIAN AUSONI AND JOHN ROGNES Definition 7.4. Let A = E(1; 2) P (t), k-1 Bk = E(0[k]) Pr(k)(t) Fp{[k]-ep | 0 < e < p} L and B = k1 Bk. Let C be the span of the remaining monomial terms in -1 E1 (S1). Then E1 (S1) = A B C in dimensions > (2p - 2). Theorem 7.5. In dimensions > (2p-2) there are subgroups eA= E(1; 2)P (v2), Bek and Ce of V (1)*T F (`; p) represented by A, Bk and C in E1 (S1), respectiv* *ely, such that (a) R* is the identity on Ae. (b) R* maps Bek+2 onto Bek for all k 1. (c) R* is zero on Be1, Be2 and eC. Q In these dimensions V (1)*T F (`; p) = eA eB eC, with Be= k1 eBk. Proof. At the level of E1 (S1), the composite map ^-1 O E1 (Rh) is the identity on A, maps Bk+2 onto Bk for all k 1 and is zero on B1, B2 and C, by 7.1 and 7.* *3. The task is to find lifts of these groups to V (1)*T F (`; p) such that R* has * *similar properties. Let eA= E(1; 2) P (v2) V (1)*T F (`; p) be the subalgebra generated by the images of the classes K1, K2 and v2 in V (1)*K(`p). Then eAlifts A and consists* * of classes in the image from V (1)*K(`p). Hence R* is the identity on Ae. By 7.1 we have C kerE1 (Rh). Thus by 7.2 there is a subgroup eCin ker(R*) ~= ker(Rh*) represented by C. Then R* is zero on eC. Note that im(R*) and ker(R*) span V (1)*T F (`; p). For by 7.1 the represent* *atives of im(R*) span A B, and the representatives of the subgroup eCin ker(R*) span C. Thus the classes in im(R*) and ker(R*) have representatives spanning E1 (S1), and therefore span all of V (1)*T F (`; p). Hence the image of R* on V (1)*T F * *(`; p) equals the image of its restriction to im(R*). Consider the subgroup B0k= Bk \ kerE1 (Rh) M r(k)-1 k-1 = E(0[k]) Pr(k-2)+epk-1(t) Fp{[k]-ep } : 0 (2p - 2) there are isomorphisms ker(R* - 1) ~= E(1; 2) P (v2) E(2) P (v2) Fp{1te | 0 < e < p} E(1) P (v2) Fp{2tep | 0 < e < p} and cok(R* - 1) ~= E(1; 2) P (v2) : Proof. By 7.5 the homomorphism R* - 1 is zero on Ae= E(1; 2) P (v2) and an isomorphism on eC. The remainder of V (1)*T F (`; p) decomposes as Y Y eB= eBk Bek k odd k even and R* takes Bek+2 to Bek for k 1, forming two sequential limit systems. Hence there is an exact sequence Y R*-1 Y 0 ! klimoddeBk! eBk- --! eBk! lim1 eBk! 0 k odd k odd k odd and a corresponding one for k even. The right derived limit vanishes since each* * eBk has finite type. Hence it remains to prove that in dimensions > (2p - 2), klimoddeBk~=E(2) P (v2) Fp{1te | 0 < e < p} and klimeveneBk~=E(1) P (v2) Fp{2tep | 0 < e < p} : Each Bek ~= Bk is a sum of (2p - 2) finite cyclic P (v2)-modules. The restrict* *ion homomorphisms R* respect this sum decomposition, and map each cyclic module 24 CHRISTIAN AUSONI AND JOHN ROGNES surjectively onto the next. Hence their limit is a sum of (2p - 2) cyclic modu* *les, and it remains to check that these are infinitekcyclic,-i.e.,1not bounded above. For k odd the `top' class 12(t)r(k)-1-ep in Bk is in dimension 2pk+1 (p-e* *). For k even the corresponding class in Bk is in dimension 2pk+1 (p - e) + 2p - 2* *p2. In both cases the dimension grows to +1 for 0 < e < p as k grows. For k odd each infinite cyclic P (v2)-module contains a class in non-negative degree with nonzero image in Be1 ~= B1, namely the classes 1te and 12te for 0 < e < p. Hence we take these as generators for limk oddBek. Likewise there * *are generators in non-negative degrees for limk eveneBkwith nonzero image in eB2~= * *B2, namely the classes 2tep and 12tep for 0 < e < p. Let e 2 ss2p-1T C(Z; p) be the image of eK 2 K2p-1(Zp), and let @ 2 ss-1 T * *C(Z; p) be the image of 1 2 ss0T F (Z; p) under @ :-1 T F (Z; p) ! T C(Z; p). We recall* * from [BM1], [BM2] the calculation of the mod p homotopy of T C(Z; p). Theorem 8.3 (B"okstedt-Madsen). V (0)*T C(Z; p) ~=E(e; @) P (v1) P (v1) Fp{eti | 0 < i < p} : Hence V (1)*T C(Z; p) ~=E(e; @) Fp{eti | 0 < i < p} : The (2p - 2)-connected map `p ! HZp induces a (2p - 1)-connected map K(`p) ! K(Zp), and thus a (2p - 1)-connected map T C(`; p) ! T C(Z; p) after p-adic completion, by [Du]. This brings us to our main theorem. Theorem 8.4. There is an isomorphism of E(1; 2) P (v2)-modules V (1)*T C(`; p)~=E(1; 2; @) P (v2) E(2) P (v2) Fp{1te | 0 < e < p} E(1) P (v2) Fp{2tep | 0 < e < p} with |1| = 2p - 1, |2| = 2p2 - 1, |v2| = 2p2 - 2, |@| = -1 and |t| = -2. Proof. This follows in dimensions > (2p - 2) from 8.2 and the exact sequence (8* *.1). It follows in dimensions (2p - 2) from 8.3 and the (2p - 1)-connected map V (1)*T C(`; p) ! V (1)*T C(Z; p). It remains to check that the module structu* *res are compatible for multiplications crossing dimension (2p - 2). The classes E(1)Fp{1te | 0 < e < p} in V (1)*T C(`; p) map to E(e)Fp{eti | 0 < i < p} in V (1)*T C(Z; p), and map by O ss to classes with the same names in the S1 homotopy fixed point spectral sequence for T HH(Z). By naturality, the given classes in V (1)*T C(`; p) map by O ss to classes with the same names in E1 (S1). Here these classes generate free E(2) P (t)-modules. For degree reasons multiplication by 1 is zero on each 1te. Hence the E(1; 2) P (v2)- module structure on the given classes is as claimed. Finally the class @ in V (1)-1 T C(`; p) is the image under the connecting h* *omo- morphism @ of the class 1 in V (1)*T F (`; p), which generates the free E(1; 2) P (v2)-module cok(R* - 1) of 8.2. Hence also the module structure on @ and 1@ is as claimed. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 25 A very important feature of this calculational result is that V (1)*T C(`; p* *) is a finitely generated free P (v2)-module. Thus T C(`; p) is a fp-spectrum of fp- type 2 in the sense of [MR]. Notice that V (1)*T F (`; p) is not a free P (v2)-* *module. On the other hand we have the following calculation for the companion functor T R(`; p) = holim n;RT HH(`)Cpn , showing that V (1)*T R(`; p) is a free but not finitely generated P (v2)-module. Theorem 8.5. There is an isomorphism of E(1; 2) P (v2)-modules V (1)*T R(`; p)~=E(1; 2) P (v2) M E(u; 2) P (v2) Fp{1te | 0 < e < p} n1 M E(u; 1) P (v2) Fp{2tep | 0 < e < p} : n1 The nth summand classes uffi1te and uffi2tep for 0 ffi 1 and 0 < e < p are detected in V (1)*T HH(`)Cpn by the classes representing uffin1te and uffin2te* *p in E1 (Cpn), respectively. We omit the proof. 9. Algebraic K-theory We are now in a position to describe the V (1)-homotopy of the algebraic K- theory spectrum of the p-completed Adams summand of connective topological K-theory, i.e., V (1)*K(`p). We use the cyclotomic trace map to largely identi* *fy it with the corresponding topological cyclic homology. Hence we will identify * *the algebraic K-theory classes K1 and K2 with their cyclotomic trace images 1 and 2, in this section. Theorem 9.1. There is an exact sequence of E(1; 2) P (v2)-modules 0 ! 2p-3Fp -! V (1)*K(`p) -trc-!V (1)*T C(`; p) ! -1 Fp ! 0 taking the degree 2p - 3 generator in 2p-3HFp to a class a 2 V (1)2p-3K(`p), and taking the class @ in V (1)-1 T C(`; p) to the degree -1 generator in -1 HFp. H* *ence V (1)*K(`p) ~= E(1; 2) P (v2) P (v2) Fp{@1; @v2; @2; @12} E(2) P (v2) Fp{1te | 0 < e < p} E(1) P (v2) Fp{2tpe | 0 < e < p} Fp{a} : Proof. By [HM1] the map `p ! HZp induces a map of horizontal cofiber sequences of p-complete spectra: K(`p)p __trc//_T C(`; p)____//_-1 HZp | | || | | || fflffl|trc fflffl| || K(Zp)p _____//_T C(Z; p)___//_-1 HZp : 26 CHRISTIAN AUSONI AND JOHN ROGNES Here V (1)*-1 HZp is Fp in degrees -1 and 2p - 2, and 0 otherwise. Clearly @ in V (1)*T C(`; p) maps to the generator in degree -1, since K(`p)p is a connec* *tive spectrum. The connecting map in V (1)-homotopy for the lower cofiber sequence takes the generator in degree (2p-2) to the nonzero class i1(@v1) in V (1)2p-3K* *(Zp). By naturality it factors through V (1)2p-3K(`p), where we let a be its image. Hence also K(`p)p is an fp-spectrum of fp-type 2. By [MR, 3.2] its mod p spectrum cohomology is finitely presented as a module over the Steenrod algebra, hence is induced up from a finite module over a finite subalgebra of the Steenr* *od algebra. In particular, K(`p) is closely related to elliptic cohomology. We conclude with some comments on the v1-Bockstein spectral sequence leading from the V (1)-homotopy of K(`p) to its V (0)-homotopy, i.e., its mod p homotop* *y. For any X, classes in the image of i1: V (0)*X ! V (1)*X are called mod p class* *es, while classes in the image of i1i0: ss*Xp ! V (1)*X are called integral classes. Lemma 9.2. The classes 1, @1, 1 and 1te for 0 < e < p are integral classes both in V (1)*K(`p) and V (1)*T C(`; p). Also @ is integral in V (1)*T C(`; p),* * while a is integral in V (1)*K(`p). The classes @2, 2, @12, 12, 12te, 2tep and 12tep for 0 < e < p are mod p classes in both V (1)*K(`p) and V (1)*T C(`; p). We are not excluding the possibility that some of the mod p classes are actu* *ally integral classes. Proof. Each v1-Bockstein fi1;rlands in a trivial group when applied to the clas* *ses @, 1, a and 1te for 0 < e < p in V (1)*K(`p) or V (1)*T C(`; p). Hence these ar* *e at least mod p classes. Since 1 maps to an element of infinite order in ss0T C(Z; p) ~=Zp and the ot* *her classes sit in odd degrees, all mod pr Bocksteins on these classes are zero. H* *ence they are integral classes. The class 1 is integral by construction, hence so i* *s the product @1. The mod p homotopy operation (P p-e)* takes 1te in integral homotopy to 2tep in mod p homotopy, for 0 < e < p. Hence these are all mod p classes, as is 2 by construction. The remaining classes listed are then products of established int* *egral and mod p classes, and are therefore mod p classes. The classes listed in this lemma generate V (1)*K(`p) and V (1)*T C(`; p) as P (v2)-modules. But v2 itself is not a mod p class. Lemma 9.3. Let x be a mod p (or integral) class of V (1)*K(`p) or V (1)*T C(`;* * p) and let t 0. Then fi1;1(vt2. x) = tvt-12i1(fi01) . x : In particular i1(fi01) . 1 = tp2 and i1(fi01) . 1 = tp12. We expect that i1(fi01) . tp2-p2 = @2 and i1(fi01) . tp2-p12 = @12, by sym- metry considerations. Proof. The v1-Bockstein fi1;1 = i1j1 acts as a derivation by [Ok]. By definiti* *on j1(v2) = fi01= [h11], which is detected as tp2 by 4.8. Clearly j1(x) = 0 for mo* *d p classes x. ALGEBRAIC K-THEORY OF TOPOLOGICAL K-THEORY 27 In V (1)* the powers vt2support nonzero differentials fi1;1(vt2) = tvt-12i1(* *fi01) for p - t. The first nonzero differential on vp2is fi1;p: Lemma 9.4. fi1;p(vp2) = [h12] 6= 0 in V (1)*. We refer to [Ra2, x4.4] for background for the following calculation. Proof. In the BP -based Adams-Novikov spectral sequence for V (0) the relation2 j1(vp2) = vp-11fi0p=pholds, where fi0p=pis the class represented by h12 + vp1-p* * h11 in degree 1 of the cobar complex. Its integral image fip=p = j0(fi0p=p) is represe* *nted by b11, and supports the Toda differential d2p-1(fip=p) = ff1fip1. This differenti* *al lifts to d2p-1(fi0p=p) = v1fip1in the Adams-Novikov spectral sequence for V (0). Conside* *r the image of fi0p=punder i1 in the Adams-Novikov spectral sequence for V (1), which* * is represented by h12 in the cobar complex. Then d2p-1(i1(fi0p=p)) = i1(v1fip1) = * *0. By sparseness and the vanishing line there are no further differentials, and i1(fi* *0p=p) = [h12] represents a nonzero element of V (1)*. Hence fi1;p(vp2) = [h12], as clai* *med. Remark 9.5. We would like to obtain the mod p homotopy groups V (0)*T C(`; p) by means of the v1-Bockstein spectral sequence. This requires, first, that we comp* *ute the product with fi1;1(v2) = i1(fi01) in the remaining V (1)-homotopy groups. N* *ext we must identify the image of fi1;p(vp2) = [h12] in V (1)*T C(`; p). Imaginably* * there is a homology differential in (4.3) on [p2 ] hitting (t)p1tp2, if (P p2)*(2) = (t)* *p1. Then most likely [h12] is detected by (t)p1tp2 in V (1)*T F (`; p), which is id* *entified under R with the class (t)p1-1 generating V (1)*T C(`; p) in this degree. The general picture appears to be complicated. References [BHM] M. B"okstedt, W.-C. Hsiang and I. Madsen, The cyclotomic trace and alge* *braic K-theory of spaces, Invent. Math. 111 (1993), 465-539. [BM1] M. B"okstedt and I. Madsen, Topological cyclic homology of the integers* *, Asterisque 226 (1994), 57-143. [BM2] M. B"okstedt and I. Madsen, Algebraic K-theory of local number fields: * *the unramified case, Prospects in topology (F. Quinn, ed.), Annals of Mathematics Stud* *ies, vol. 138, Princeton University Press, 1995, pp. 28-57. [Bo] A.K. Bousfield, The localization of spectra with respect to homology, T* *opology 18 (1979), 257-281. [Br] R.R. Bruner, The Adams spectral sequence of H1 ring spectra, H1 Ring * *Spectra and their Applications, Lecture Notes in Math., vol. 1176, Springer-Verlag,* * 1968, pp. 169- 214. [DH] E.S. Devinatz and M.J. Hopkins, Homotopy fixed point spectra for closed* * subgroups of the Morava stabilizer groups, preprint (1999). [Du] B.I. Dundas, Relative K-theory and topological cyclic homology, Acta Ma* *th. 179 (1997), 223-242. [EKMM] A.D. Elmendorff, I. Kriz, M.A. Mandell and J.P. May, Rings, modules, an* *d algebras in stable homotopy theory, Mathematical Surveys and Monographs, vol. 47, A* *mer. Math. Soc., 1997. [GH] P.G. Goerss and M.J. Hopkins, Realizing commutative ring spectra as E1 * * ring spectra, preprint (2000). [HM1] L. Hesselholt and I. Madsen, On the K-theory of finite algebras over Wi* *tt vectors of perfect fields, Topology 36 (1997), 29-101. [HM2] L. Hesselholt and I. Madsen, On the K-theory of local fields, preprint * *(1999). 28 CHRISTIAN AUSONI AND JOHN ROGNES [HSS] M. Hovey, B.E. Shipley and J.H. Smith, Symmetric spectra, J. Amer. Math* *. Soc. 13 (2000), 149-208. [Li] S. Lichtenbaum, On the values of zeta and L-functions: I, Ann. Math. 9* *6 (1972), 338-360. [Ly] M.G. Lydakis, Smash products and -spaces, Math. Proc. Camb. Philos. Soc* *. 126 (1999), 311-328. [MR] M. Mahowald and C. Rezk, Brown-Comenetz duality and the Adams spectral * *sequence, Amer. J. Math. 121 (1999), 1153-1177. [MS1] J.E. McClure and R.E. Staffeldt, On the topological Hochschild homology* * of bu. I, Amer. J. Math. 115 (1993), 1-45. [MS2] J.E. McClure and R.E. Staffeldt, The chromatic convergence theorem and * *a tower in algebraic K-theory, Proc. Amer. Math. Soc. 118 (1993), 1005-1012. [Mo] J. Morava, Noetherian localisations of categories of cobordism comodule* *s, Ann. Math. 121 (1985), 1-39. [Ok] S. Oka, Multiplicative structure of finite ring spectra and stable homo* *topy of spheres, Al- gebraic topology, Proc. Conf., Aarhus 1982, Lecture Notes in Math., vol* *. 1051, Springer- Verlag, 1984, pp. 418-441. [St] M. Steinberger, Homology operations for H1 and Hn ring spectra, H1 Ri* *ng Spectra and their Applications, Lecture Notes in Math., vol. 1176, Springer-Ver* *lag, 1986, pp. 56- 87. [Qu] D. Quillen, Higher algebraic K-theory, Proc. Intern. Congress Math., Va* *ncouver, 1974, vol. I, Canad. Math. Soc., 1975, pp. 171-176. [Ra1] D.C. Ravenel, Localization with respect to certain periodic homology th* *eories, Amer. J. Math. 106 (1984), 351-414. [Ra2] D.C. Ravenel, Complex cobordism and stable homotopy groups of spheres, * *Pure and Applied Mathematics, vol. 121, Academic Press, 1986. [Ra3] D.C. Ravenel, Nilpotence and periodicity in stable homotopy theory, Ann* *als of Mathe- matics Studies, vol. 128, Princeton University Press, 1992. [Re] C. Rezk, Notes on the Hopkins-Miller theorem, Homotopy theory via algeb* *raic geometry and group representations, Contemp. Math., vol. 220, Amer. Math. Soc., * *1998, pp. 313- 366. [Ro1] J. Rognes, Two-primary algebraic K-theory of pointed spaces, preprint (* *1998). [Ro2] J. Rognes, Etale maps of S-algebras, in preparation. [RW] J. Rognes and C.A. Weibel, Two-primary algebraic K-theory of rings of i* *ntegers in number fields, J. Amer. Math. Soc. 13 (2000), 1-54. [Ts] S. Tsalidis, Topological Hochschild homology and the homotopy descent p* *roblem, Topol- ogy 37 (1998), 913-934. [V] V. Voevodsky, The Milnor conjecture, preprint (1996). [Wa1] F. Waldhausen, Algebraic K-theory of spaces, a manifold approach, Curre* *nt Trends in Algebraic Topology, CMS Conf. Proc. 2, Part 1, 1982, pp. 141-184. [Wa2] F. Waldhausen, Algebraic K-theory of spaces, localization, and the chro* *matic filtration of stable homotopy, Algebraic topology, Lect. Notes Math., vol. 1051, S* *pringer-Verlag, 1984, pp. 173-195. Department of Mathematics, ETH Z"urich, Switzerland E-mail address: ausoni@math.ethz.ch Department of Mathematics, University of Oslo, Norway E-mail address: rognes@math.uio.no