subsequent induction of p34" bur p34 expression was not required for induction of emyb oF c-myc, the inhibition of p34°" expression with antisense oligo- mers didnot atfect induction of either proto-oncogene. Thus, p34 expression ‘ceurs relatively late in the sequence of Gr-S control genes, subsequent to the induetion of e-mph and cmye. These reslts suggest that some of the known requirement for ¢- mye and comyb for the Gr transition in lectn-stimalated T cells (16) could be medl- ated through induction of p34, because reduction of p34°* expression, either di rectly by antisense «2 oligomers or ind: rectly by blocking either cmye or cmyb induction, produces equivalent inhibition of DNA synthesis. The observed degree of inhibition of p34“ expression after treat- iment of cells with antisense mph oF mye could, by itself, be suficient to block cells at Gy-S. However, emyh and eomye ikely regu Inte the expression of many genes in Gy and ray control several parallel events that are ako requieed for the Gy-S transition. ‘Our antisense oligomer experiments in cate that p34? is not likely to have a regulatory function at Gy-Gy in lectin-act vated T cells. The small amount of p34° present in resting T cells may be characteris- tic of Gy cells and a similar finding has also been reported for serum-starved fibroblasts (A) and quiescent baby tat kidney eels (23). It is possible, however, that p34" plays some role in Gp cells, because expression of 2 has been described in nonproliferating brain eels (22) Finally, i is worth considering that aber= rant expression of camye and canyh in hema- topoietc cells can eesule i leukemia and that loss of expression of Rb is associated with transformation of many cll lineages, inclu ing T cell leukemia (24). It is therefore possible thar p34 itself of regulators of p34 function at Gy'S could contribute 0 transformation of hematopoietic cel [REFERENCES AND NOTES 1. D. Asma il 8,371 (1988; J. Gaia, fl 88 43 (1986) ]-C- Labbe cal, Nata 398, 31 1988) 2G Dra and D, Beach, Cal 4,17 (1988. 3G Draco ay Notre 396, 738 (1988), V. ‘Sinan an P Nano Cl 45,261 (1986) MCG Lav eral, Nee 833, 67 (988), 1. Danphy ad. Newport a8, 11 (1989). Gautier T” Mattar F- Nase, | Mole, Nanve $39, 626 (1989); K- Gould and P Non, hi 342, 539,989) AO, Moris G. Drea D. exh, 1. J. Wang, Cal $6, 193 (1989) 6, P Nunc and ¥- ise, Nut 292,558 (1981); S Reeds J A. Hager, AT. Lorine, Pre Na ‘al Si USA, By 4088 (1988) 2, A Giowdan eal, Ci 88, 981 (1989) 8K Rutowol asi 87, 393 1989) 9. S.A Cannstes snd). Gin, Som Homa 25 153 (94th, TT. towel Ng J. Na. 2 717 1974), 10, G. Crate, sine 243,355 (1989) 820K Ban Hl teagan coming the aman gone wa exci fom FORD 26 aise ‘Stitt poner of ps2 (Braker np lid data to generate pSAFI, phiDS2 wa der te of poct (RM. Myer, LS. Leman, T Mani Situ” 229, 242 1985). pSAFI win Sigoed with Ba {and Pv Tl and the 0940 femene of amano? as wae at probe for ‘Note bring afer being wth P(A. Fe Fenberg and 8. Vogcicin, Aa! Bosh 133, (ona 12, 4 2350 protcin agent ercompasing the COOH rms of p34 epning at mei tine ride 8) was ape bata Ge the I pomater, Pcie war olted by SSP amide ge ekcropore and eed moae Now Zealand Whte abba (Mt Lee and HP rca Worms: unpubised di, Purged pi ws bled fom an overepresing Sain of Edens ots dered [rexel Dra, D. Beach, EMBO} 3507 (198) ad Couple to cranogen bromides Sephtone S88 mg pr milter et (Paracay. 1M, $2080 fan TE fap containing the baman {42 gone wae fom POH (i) an ere Sto be Ram Hi st ofthe bacon eee Son vector VISIT TA. Vi Lichow and BED, Sunes B/Tedoiay 6,47 (1988), Reco ant vw cong pa plague pied and {Sadao mie Spee agp) ce Al provslrereatng © val Popagaon, Slaon, Sn infection were pero eel ay de bed (M.D, Sune and. ES 9 No °F Mo fr vis Vers nd il Cr ne) (Tex Agra Experiment Staton Bilin No. 1855, Colge Strom, 1, 1987) Fe percent of toa all roten wasp che ujnay of which was okie an unpexponyt fat Parker and Pc Worna, unpaid 15, Pe Kui, J dma 138, 4114 1987) 16, RH ay Nome 228,448 (1987) A. Gove a Sout 248, 190° (1989). AM. Gira ah) egy Ned 170.108 0989) 171A, DeCapeio a, Cal $8, 1085 (1989): Bochkonch A ul, E Haske ip 1097, DAL. Chen al sp 10S, K Mbars ea Sie 246, 150 (1089) He ea Ones 4307 (1989), 18 Y Farawa a, Pe Nah Ae So USA. 87, m0) 19. Y Tapa ea Bio, Bp. es. Comma. 164, S580 1989). 20, 1A. DeCapio ta, Cal $4,275 (1988); J. W. Talo ey i $6, 57 (1989) J. Low sin. 60, 347 (1990) 21, K’ Kaley, 8. HL Cachan, C.D. Sales, P. Leder, ii 35, 6031988) TC Recd).C. Alpers, P-C Nowell RG. Hoover Pre Na Aand SUS. A 85,3982 (1986), A Shipp and E 1- Rember, nuns. 139.2148 (1987), 22, J. Stem and Ke AL Sith (ie). 28, G. Drcta, D. Beach, E. Moran, Onagne 2,583 (198). 24. J. Cheng a, 28 ¥ Gardasil Siem 288, 208 78, 730 (1990) Res 47,2889 (1987). 26, M,G Lec and P. Nurse, Noo 327.31 (1987), YF. VA wacker al, MAL mat 26, 1037 0989). 28. | ine and. Fler, Col 88, 83 (1089). 28. ELL Wacktrom eal Pac Nat Aad Sa U.S.A 85, 028 (1988). Holy RL Redes, AW ‘ren Mol Ci. Bal 8,568 (988). Werthank J. DeCapeo and B- Dear foe etal reve ofthe data andthe manuspesG- Drees for he bacesal plo orepeatucery Nurse for POBDSL; and Morimoto forthe monoconal thuisdy eo CDS. Seppared in prey PTS pants CASSIG7, CAAPBIS. CASD7O7 and CASHIRS XY is supported in art by 2 felowstip fem the shar Memo Foundation and the Mochi Foundation for Medical and Dhamaeutea Re Searcy snd H.-W by3 Whitaker Hes Sensex Fn! Award JDK ea Scolar of the Lenker Sosy of Ames 8 May 1990, acepted 29 Ang 1990 Female Preference Predates the Evolution of the Sword in Swordtail Fish ALEXANDRA L. BaSOLo* ‘The study of female preferences and the evolution of male traits has until recently centered on genetic coevolutionary mechanisms. An alternative mechanism posits that 1 preference results from a preestablished bias in the female information-processing, system: Jing from sources independent of sexual selection, Male traits then arise that are selected by this preexisting preference. The genus Xiphophorus consists of swordless platyfish and swordtails. Swordlessness is the primitive state. In this study, female platyfish, X. maculatus, were found to prefer conspecific males with artificial swords ‘over those without swords, despite evidence that the common ancestor of platyfish and swordtails was swordless. These results suggest that the evolution of the sword in the ‘swordtail clade was a consequence of selection arising from a preexisting bias. posed by Darwin (1) as a mechanism to explain the evolution of elaborate traits in males that appear to decrease their ‘survival. One typeof sexual selection, female choice, involves a preference by females for traits in males. Until recently, models that inves of Tas Aust, TX 78712, “Pesce adden: Department of ological Senex, nero Clon, So tars, 93106 stress the coevolution ofa female preference and a male trait have dominated theoretical and empirical treatments (2, 3). Few data have been produced that support one of these models to the exclusion of another, and alternative, testable models have been proposed [see (4) for review]. A non-coevo lutionary explanation for the evolution of a female preference and a male trait proposes that biases in the female sensory or cognitive system, or both, arse and increase t0 a high SCIENCE, VoL. 250,frequency, or fixation, for reasons indepen- dent of sexta selection (5-9). Subsequently, the evolution of male traits occurs when @ ‘newly arisen trait “exploits” preesisting bias- cin females. Thus, a preexisting bias in the female information processing system, re- sulting from evolutionary processes. inde- pendent of sexual selection, is present before the appearance of the male trait. Unlike ‘models based on coevolutionary processes, a preexisting bias model does not requite fe ale preference to evolve because of sexual selection, nor does it requite the coevolution Of the preference and trait. However, after the male trait arises and is selected by the preexisting bias, other sources of selection may act in conjunction to further elaborate ‘or modify the expression of the preference and the trait To demonstrate that preexisting biases have resulted in the evolution of a male trait, three criteria should be met. First, it must be established that in species in which the tait js present, there is female choice based on variation (either natural or artificial) in the trait (10). Second, it must be established that the absence Of the male trait is the ancestral state within a clade, Third, it should be established that in a species hav- ing the primitive state, females prefer the male trait even though it was not generally present in the evolutionary history of the species, In this report, I address these erte- +a and present data that are consistent with 4 preexisting bias model forthe evolution of the sword in the genus Xiphophons Xiphophons sa genus of smal, freshwater fish in the live-bearing family Poccilidae of Central America, The genus includes the platyfsh, which are primarily swordless, and the swordtails, males of which typically have 4 sword, a colored extension of the lower ‘margin of the caudal fin (11, 12). The sword does not develop until sexual maturity, at Which time the rays at the base of the caudal fin begin to lengthen in males and acquire a border of black coloration. Swordlessness is ‘considered to be the ancestral state for the genus (Fig. 1). 1 have shown that female green swordtajls, X. eller, prefer males with longer swords (13). The preexisting bias ‘model prediets chat male traits arise and are favored by preestablished preferences. IF this is true, females in a closely related, swordless species such as X. maculatus, the southern platyfsh (14), should also have a preference for a swordlike structure, even though the ancestral state of swordlessness is main- tained. This hypothesis was tested through female choice tests to determine whether preference for a swordlike structure is a primitive stare within the genus Xiphophons Choice tests were conducted in an aquari- tum (45 em by 41 em by 87.5 em) divided by 9 NOVEMBER 1990 Table 1. Diference in time females spent with sales of dtferent sword statue (SE in parenthe ss). "Yellow" represents males with areca, 24 fm, surgically artached yellow plastic swords “Transparent” represents sales with 24-mm, sur gical atachod dear partic swords simulating the absence ofa sword! Probabilities (P) ae rom Wilkoxon matched-pats signed-rank test ‘Mean time Body @ se ——— op (um) Yellow Te parent 35 610125) 478126) 16 0.039 SI 685 (39) 44247) 14 OOLL 33 694 (67) 414 (68) 170081 34750 (41) 375 44) 140.002 28 732 (39) 395(28) 90.008 34710054) 411153) 140.026 ‘wo Plexiglas plates into hree equal com- partments (13). The middle compartment Was divided into three sections: one center section and a section adjacent to each of the outer compartments. A test female was ped in an opaque tube suspended from above the center section, and a test male was placed in each of the outer compartments for a 10-min acclimation period. A test trial was initiated by raising the tube and releas- ing the female into the center section, begin- ring a 10-min observation period During the observation period, the be- havior ofthe test subjects and the amount of time the female spent in each of the sections adjacent to the male compartments. were recorded. After this period, the female was ‘once more placed in the tube and the males ‘were switched from one end compartment to the other (10 control for side bias), and a second acclimation and observation period was conducted. During the choice tests, ‘ales and females actively courted one an” ‘other. The time of the two observation periods was summed, and a Wileoxon ‘matched-pairs signed-rank test was used on the data set for each male-female combina tion to compare the difference in time spent with each of the males. Proximity, coupled with active courtship, was used as an indica tor of female mating preferences; mating was not observed directly (15). Test subjects were either lab stock or progeny of wit ‘caught fish (16). Paired males were matched for body sie (within I-mm standard length) and body coloration, I tested female preference for a sword structure in the swordless X, maciatus by surgically applying a plastic sword to platy- fish males (17) and conducting, female choice tests. Males were matched for body size and coloration. Paired males each re- ceived a 25-mm sword; one male received a yellow sword with a black border, the other 4 transparent sword (18). The transparent ‘sword simulated the visual absence of a sword, while controling for surgery effets and influences of the artifical sword on ‘swimming behavior. Males were tested for female preference after a 3-day postoperative period, Males who underwent this surgery recovered and performed normal courtship behavior during female choice tests. Six pairs of males were tested, each with 9 to 16 females. In all six sets'of tests, females preferred the male with the yellow sword to the male with the transparent sword (G0 tailed binomial P = 0.032; Table 1), ‘An additional experiment, a sword rever- sal test, was conducted in which female preference was observed for a male with a yellow sword over a male with a transparent ‘sword (Wilcoxon matched-pairs signed-rank test: n= 14, P= 0.03), and then the origi- nal swords were removed and replaced with the other type of sword. With this reversal in sword status, female preferences were re- versed (Wilcoxon matched:-pairs signed- rank test: «= 14, P= 0.01); females dem- onstrated a preference for the male with a yellow sword, although they had previously discriminated against him when he had a transparent sword, These data suggest that the females were basing their choice on sword presence and not other traits. ‘The absence of a male trait a the ancestral condition has been demonstrated in the frog genus Physalaemus (7). Physalaemus pustlosis females prefer male chucks at the low fre: quencies they detect best. In the closely related species, P. coloradorim, in which Prapeta Praysh Prati wort Fg. 1, Abbreviated phylogeny ofthe genus Xi plophone (1,1) (sone akdonal species ave fech decrt be phylogenci nas has oe teen published). The plays are generally une worded, and most spice of words postese 9 Svond (Is Swordesnes appear 10 be the Primitive stat ste ster group, Priel does Aeron a svon and he swe cosine oF {Colored eatnson a the ventral agin of the Caudal in, a synapomoephy of the sword pis (9). Morphologic cetrophoret, pig oti and reigns charters wor el to cash the plivlogeny Adina, pacers ‘odin were considered in some ces {te cine vo which the playfsh X. manne Eelonge designated by asters, *) REPORTS. 809males do not produce chucks, female neuro physiology also suggests that the inner ear is tuned to lower frequency calls. Therefore, female biases in the inner ear appear to have evolved before the prostuction of low-fre quency chucks in males. The present study with playfish demonstrates that females ex hibit a preexisting bias for a trait not gener ally present in the evolutionary history of the lineage, The three criteria necessary to ‘demonstrate clearly that preexisting biases ‘ean result in the evolution of male traits have bbeen met with this study: (i) there is varia tion for sword length in. helleri, the green sword, and females prefer males with longer swords (13); (i) the phylogeny for Xiphophorus suggests that the ancestral con ition is swordlessness (11, 19) (Fig. 1); and (ii) in the swordless platyfish, X. maculatus, females demonstrate a preference for males with swords. Obviously, only traits chat ie within the dlecection limits of the female sensory anc Processing systems can be selected via female choice; thus, the potential range of sexually selected male traits i constrained, In preex- isting bias models of sexual selection, the bases for the initial bias are numerous. A swore may increase the perceived boxy size cof a male (8); thus, a preexisting preference for large males could, a a by-product, result in selection for the sword once the trait arises in a population, Similarly, a preexs ing preference for rare, novel male types could intially select a swordlike seructure; however, once an initially rare, novel trait becomes’ common, some other mechanism of preference or source of selection would be necessary t0 drive the trait 10 a high frequency oF fixation. Therefore, preference for novelty per se cannot explain the evolu- tion of the sword in swordtal fish Female sensory or cognitive ystems could be initia ly biased in response to environmental sources of selection. For example, a female ‘may be adapted to recognize a moving shape in the environment because that moving shape resembles a favorite food, and females have developed a search image for it. This bias in the female detection system could subsequently bias the direction and nature ‘of sexual selection if a mutation for males resembling this moving shape arises. In ad- dition, new traits that increase conspicuous- ness could lead to selection for the trait. All ‘of these possibilities share one common factor: the presence of the bias that results in a preference precedes the appearance of the male trait. In contrast, a cocvolutionary mechanism requires the joint evolution of the trait and ehe preference AA less parsimonious explanation for the preference of X. maculatus females for sword ‘ed males is that this preference is derived separately within the X. maculae clade. Such a preference may be for novelty per se for some other characteristic, However, this interpretation of the results requires ewo independent evolutionary events: the evolu- tion of a preference in the platyfish lade and the evolution of a sword preference in the swordtail clade. A preesisting bias hypothe- sis, im which a single evolutionary vent accounts for a shared preference in platyfsh and swordtails, is thus a more parsimonious explanation, ‘This work suggests that the common an: cestor of the unsworded X. maculatus a the sword X, heller possessed a preference for a swordlike structure. Since phylogenetic hypotheses suggest chat males of this com: ‘mon ancestor were unsworded, the preexist: ing bias resulting in a sword preference appears to have evolved before the sword itself; the sword thus may have subsequently arisen in the swordeal clade and been select: ed for by this preexisting bias. Models stress: ing coevolutionary processes eannot explain ‘evolution of a female preference in the ab sence of a male trait, and, therefore, a preex: isting bias model seems to best explain the evolution of the preference and the initial, evolution of the sword. This conclusion does not preclude a role for coevolutionary processes. Once the sword has evolved, the initial bias and the trait could potentially be ‘modified by other processes. However, the Xiphophoris system provides evidence for preexisting bias models of sexual selection; a role for other processes remains speculative, "REFERENCES AND NOTES, 1. G. Darwin, On the Origin of Spats by Mam of ‘Natl Selon ote Proc Fare Rats the Sige Le (hn Morey, London, 1889): “The Dos of Stn, an Stony Relat Sex (an Mares, Lenton, 1871), 2. A Zaha, Thor “Bot 3, 205 (1978). ‘ever ia Sc Seton andthe tof a icieiinr, 8. CampéB. (Akine Chica, 1983), 9p. 136-179 A. Kode Brown ant) TE Bron. Ar 124, 309 (1988) Proponents of “gon” gencs sugges that one oe aie tits {fe acd saan cea ner af al toes and that female perce eve aden sect fot female 10 ete wth mes who have genomes "Svored by ome sour of mur scion, 3, JOA Fier, The Conta! Thor of Nol Se vk, 1930, 1958): Lae. Pre U'S'A.78°3721 (198i): P-O'Den Mate Che ate, El (Cambie ‘ress, Cambs, 1983, pp. 83-66 Fa ian runaway slconsts sees the portance of linkage sya enveer fra preferences nt stl Slected make ate: prceenes eve {1 nlated response o mae tes svn. 4M Kiar A Ror El St 18,43 (1987), fn Scxul‘Scon Teng the Aeon JW Bradbury and MB, Andersson, El (Wey Chuch ‘ster, United King, 1987), pp. 7-82 {GL Bast Hw Animal! Comma, TA Scheak, Ba nina Un Pre Bln, 1977) p TBI: Buse Avian igo. PA ray and D. W. Mock, Eds, [Ontong 7 (1985), pp 22-44 6 LA-Balerand T, MeLeln, Anu Re Et Sy 2». 19, 395 (1988) JA. Ele, in Speci nd rg, D. One and JA. Eadie, Eds (Si. duce, Sunderland, MEA, 198), pp. 625-648, MCT. Roan Ase 3436 (1990) 0). Rye tia Ran, Eno 44, 308 (1990), MOP Ryan Ox Sire Ele, npr DPreaitng base m the eae cay ter ave been terol odk sersory dive (6) and smsory exploitation (78); Both of seers are describe {8 isin the dcton of evolu that se the sensor stem af 2 spose Howeter, sory five spies tos fore by nara ton, ‘ul secon, or combination ofboth and en tect the drecton of evolu tn the absence of fehl eetion Sensory elon pps pei Esto seul selected ei However, abence of 3 prerence does ne phy that's prckrence hae ener een presen am te fvolutnary history of the specs apd hus hat, Emule cote never pay se the evolion of thet MM Rochenberper oa, Am Mac Now 2975 ios) carter generally shar by all swords is ‘elope solraton on the venel arg oe ual in Howeser, ts carson 1s absent in three mothe oda (i AT, Bacio, Anon eh. 40, 832 1990), {Wik X mnt never devo an eenaon of dhe abl Gin, bu hey have Seen expecta ce to preces swoed (9), Ariel selene is results fh wh eangeted upper and leer mari ofthe cal fy ac bese some Pate specie. The males of two species of Pay sh Xie and. aks, develop an Unpigncnted extension (10 8 mn) ofthe lower tgif the cl fn (1. 1) Avena ee “Tse sabes iter viginated om hb stock pop lnsons mained by Ke Kallman, Osborn Ge vine 256, 595 recs Laboratory, New Yook Aquarium, or were progeny fea Eedeangh nti rn che Be TeetRier Bctos coleted in 1986 and LORS, Lab ‘tock sale al eer te eeu wipe, ‘lacus Bly coloraon and ed ses oe haa Specie ody (Sp) slr ic, Blinc males were Sntched for an of s mamber of known olor = Before appticauon of the svord, the subject was sec th MS 88, Sopa nape oe Terang the amet ofthe celal potoncle {hth 2.0 stare nce tie to secre he patie Sond, Tyee dole revened spel Knot ere {wed to ne off the sure The mes By as fe in Nowagu and SirssCou during the 40 50s shat he procedure segue and Wounds Ws pled to the sures sere. cnc the pony that the Gemaes were exon the trneparen sword 3 tee ts CON {Ried in which preerence fora male with a yell fond Was estaisbed, and then the transparent ‘ond wae cut back sla Ibsen nay hat da at ‘Sten jon the dal elge of the cl fi A Stee sof ele choice fer wat then enc fT ds ts peeencs id pr change There teas aloo signe difrence inthe tine spt tthe me wth the ranepacet sword or the ‘elite cine spent wh these wth he ew Soest rs ths marpultion (A. Bas, ‘apobie ts) DBs, Ronn, Ba, Am Mu. Nit. Hie, 162, 267 (197, Support by NSE Doctoral Disertion Inpro tat gram BSR 8700944, 4 Theore Rosset frank the American Museum of Natal ior, Shas Raney Avan Ameran Soy of Ichthyol fas and Herpewiogsts [dusk KD. Kallman for vice a fh the Mansy of Aiur Bele for providing oleeeng permits JJ Bull, Kinparick and. Pee Yor providing rca me} Gartner for esearch sane and Barker. Bul JA Eder, R- Gomulkewez, K Kalman, M-Kiskatik, ME Ryan, W Wage, ant thee anoayine vevicwen fer consuctve itr ‘am ofthe csprmental dengn sr the names 27 Mar 1090; secepta 11 July 1990 SCIENCE, VOL. 250,