Flying Apart : Mating Behavior and Speciation

Transcription

1 Flying Apart : Mating Behavior and Speciation CHRISTINE R. B. BOAKE Bi o d i v e rsity is the outcome of the processes of s pecies form a ti on and ex ti n cti on. S pecies form a ti on ( s pec i a ti on) is by far the lesser unders tood of those two proce s s e s, but it is clear that the heart of s pec i a ti on is the evo luti on of reprodu ctive isolati on, that is, the failu re of genes to move freely bet ween two species (Mayr 1963, Ha rri s on 1998). Wh en genes do not move freely, e ach s pecies is an indepen dent evo luti on a ry unit. In animals, t h e evo luti on of reprodu ctive isolati on is usu a lly gradu a l, i n con trast to plants, in wh i ch reprodu ctive isolati on can ari s e in on ly one gen era ti on thro u gh po lyp l oi dy, t h e mu l ti p l i c a ti on of s ets of ch rom o s om e s. Reprodu ctive i s o l a ti on can invo lve factors that affect ei t h er the prob a bi l i ty of z ygo te form a ti on (pre z ygo tic isolati on) or the vi a bi l i ty or reprodu ctive su ccess of hybrids (po s t z ygo ti c i s o l a ti on ). In this arti cl e, I focus on the form er, ex p l ori n g a pproaches to stu dying the evo luti on of pre z ygo ti c reprodu ctive isolati on in animals. Reprodu ctive isolati on can evo lve as a byprodu ct of ecol ogical diver gen ce or as a re s ponse to direct sel ecti on on reprodu cti on (for a revi ew of these top i c s, s ee Sch luter 1998 ).E co l ogical diver gen ce is the process of ad a pt a ti on of d i f ferent pop u l a ti ons to different envi ron m en t s. Ei t h er pre - or po s t z ygo tic reprodu ctive isolati on may fo ll ow this divergen ce : Pop u l a ti ons may not exch a n ge genes because they h ave ad a pted to different nich e s, or pop u l a ti ons in new envi ron m ents may persist long en o u gh for reprodu ctive i s o l a ti on to evo lve because of the gradual acc u mu l a ti on of gen etic differen ce s, rega rdless of ad a pt a ti on. Di rect sel ecti on on reprodu cti on wh i ch can invo lve anything from l on g - d i s t a n ce attracti on to co u rtship beh avi or to s perm egg interacti ons (We s t - E berh a rd 1983, Pa lu m bi 1998 ) gen era lly produ ces pre z ygo tic isolati on. As a factor in reprodu ctive isolati on, eco l ogical divergen ce has received far more atten ti on than beh avi oral and s perm egg isolati on, perhaps because until recen t ly scientists foc u s ed more on how species maintain their disti n ctness than on how species form (Fo s ter et al ). Moreover, eco l ogical diver gen ce can be stu d i ed well by examining species that have alre ady re ach ed su b s t a n ti a l reprodu ctive isolati on, wh ereas the qu e s ti on of h ow spec i e s form requ i res a stu dy of pop u l a ti ons that are in the proce s s of d iver gi n g. D a rwin re a l i zed that oceanic islands are parti c u l a rly va lu a ble for the stu dy of s pec i a ti on. The waifs that manage IF S E X UA L S E L E C T I O N D R I V E S S PE C I AT I O N, I T S H O U L D B E P O S S I B L E TO F I N D T RA I TS T H AT A R E BOT H S E X UA L LY S E L E C T E D A N D I N VO LV E D I N B E H AV I O RA L I S O LAT I O N to re ach isolated arch i pel a goes and find a favora ble habi t a t a re in a situ a ti on that perm i t s, and may prom o te, s pec i e s form a ti on. The Galapagos and Hawaiian islands thus are excell ent sites for studies of s pec i a ti on (Wa gn er and Fu n k , Grant and Grant 1996), a l t h o u gh other arch i pel a goe s also are important (Cl a rke and Grant 1996). O f the many groups of plants and animals whose rad i a ti on has been s tu d i ed in Haw a i i, the Dro sop h i l a in parti c u l a r, because of t h eir divers i ty and el a bora te mating displays (Ri n go 1977), h ave sti mu l a ted re s e a rch on the hypothesis that mati n g beh avi or can drive the process of s pec i a ti on (Ka n e s h i ro ). S exual sel e ction and spe ci a ti o n Sexual sel ecti on is that part of m a ting beh avi or that concerns mating su ccess (Darwin 1871); it drives the evo luti on of ex trava gant male orn a m ents su ch as plu m a ge, c a ll s, ph erom on e s, and antlers. These orn a m ents are usu a lly disadva n t a geous to a male s vi a bi l i ty or su rviva l, but they are adva n t a geous in that they increase a male s mating su cce s s. In most spec i e s, males have a high va ri a n ce in mating su c- cess and thus are the sex u a lly sel ected sex, wh ereas fem a l e s h ave a uniform ly high prob a bi l i ty of m a ting and are ra rely s ex u a lly sel ected. Sexual sel ecti on can invo lve both intermale com peti ti on for access to mates and female ch oi ce a m ong po ten tial mate s ; the rel a tive import a n ce of these two processes differs among spec i e s. Formal gen etic models dem on s tra te that sexual sel ecti on can lead to spec i a ti on wh et h er or not eco l ogical diver gen ce Christine R. B. Boake ( cboake@utk.edu) is a professor and associate head of the Department of Ecology and Evolutionary B i o l o gy, U n i v e rsity of Te n n e s s e e, K n o x v i l l e, TN American Institute of Biological Sciences. June 2000 / Vol.50 No. 6 BioScience 501

2 Fi g u re 1. A co m pa ri son of Dro s oph i l a m a l e s. The two l a rge flies are Dro s ophila silve s tri s ( l eft; round head) and Dro s ophila heteron eu ra ( m i d dle; broad hea d ), wh i ch are s h own with a Dro s ophila mel a n oga s ter for size co m pa ri so n. Ph oto: Todd Ca m pbell. is pre s en t. In a qu a n ti t a tive gen etic model of an idea developed by Fisher (1958), L a n de (1981) showed that even in the absen ce of eco l ogical diver gen ce, female mating preferen ces can alter male displays to su ch a point that the pop u- l a ti on can be con s i dered a new spec i e s. This model assu m e s that both female preferen ces and male displays are contro ll ed by several gen e s. Females are most likely to mate with males that they prefer; t hu s, t h eir daugh ters wi ll show a preferen ce similar to their own and their sons wi ll have a d i s p l ay similar to the father s. Ad d i ti on a lly and cri ti c a l- ly d a u gh ters carry the genes for the father s display type and sons carry the genes for the mother s preferen ce (i.e., t h ere is a gen etic correl a ti on bet ween the preferen ce and the tra i t ). Because the preferen ces and the traits are correl a ted, a sel f - rei n forcing gen etic sys tem is set up and the pop u l a- ti on can run aw ay to mating beh avi or that is very different from the starting poi n t. The en d point can be so different that beh avi oral isolati on, and thus a new spec i e s, evo lve s. Al t h o u gh Lande s ori ginal ru n aw ay model did not ad d ress eco l ogical factors (Lande 1981), L a n de and Ki rkp a tri ck later mod i f i ed the model to show that eco l ogi c a l d iver gen ce bet ween pop u l a ti ons can be accom p a n i ed by d iver gen ce in female mating preferen ces (Lande and Ki rkp a tri ck 1988). Perhaps the most su rprising re sults com e f rom a later model gen era ted by Tu rn er and Bu rrows ( ), wh i ch also is based on the ru n aw ay proce s s. Th e s e re s e a rch ers showed that sym p a tric spec i a ti on (spec i a ti on with no geogra phic barri er to gene exch a n ge) can be driven by female mating preferen ce s. Th ey found that if fem a l e s a re su f f i c i en t ly ch oo s y, sampling many males before accepting a mate, s el ecti on can be strong en o u gh to prevent gen e exch a n ge bet ween sym p a tric pop u l a ti ons with differi n g types of d i s p l ays. Before this model was pre s en ted, t h e process of s ym p a tric spec i a ti on had been thought to be ra re at best (Futuyma and Mayer 1980). The recent model su g- ge s t s,h owever, that sym p a tric spec i a ti on is far more plausible than many re s e a rch ers had imagi n ed. S c i en tists have also used non m a t h em a tical arguments to rel a te sexual sel ecti on to spec i a ti on. For ex a m p l e, Ri n go (1977) propo s ed that the con comitant evo luti on of a new male display and female preferen ce could re sult in reprodu ctive isolati on, wh i ch could explain the great divers i ty of s econ d a ry sexual ch a racters among species of Haw a i i a n Dro sop h i l a. As a re sult of a stu dy of four species of Haw a i- ian pictu re - wi n ged Dro sop h i l a, a n o t h er re s e a rch er, Ka n e- s h i ro (1976, ), devel oped the hypothesis that spec i a ti on in this group has been driven by sexual sel ecti on ; he propo s ed that females in more derived pop u l a ti ons would be less discri m i n a ting in their ch oi ce of males than females in a n ce s tral pop u l a ti ons bec a u s e, in a small pop u l a ti on after a fo u n der even t, females that were high ly discri m i n a ti n g m i ght not mate. We s t - E berh a rd (1983) pre s en ted det a i l ed a r g u m ents in su pport of the hypothesis that differen ce s bet ween animal species in mating beh avi or could be attri b- uted to sexual sel ecti on. Her re a s oning inclu ded the poi n t s that sexual sel ecti on is of ten both strong and directi on a l and that the ru n aw ay process could be initi a ted by many gen eti c, beh avi ora l, and eco l ogical factors. She propo s ed that sexual sel ecti on could have driven spec i a ti on in many animal gro u p s,i n cluding bi rd s, i n s ect s,c ra b s, and lizard s. Both the gen etic models and the verbal models propo s e that female mating preferen ces can ch a n ge the mati n g beh avi or of a pop u l a ti on so gre a t ly that reprodu ctive isolati on can evo lve. This propo s i ti on su ggests a hypothesis for ex peri m en t a ti on n a m ely, that sexual sel ecti on and spec i a- ti on form a con ti nu u m, or, in other word s, both sex u a l s el ecti on and beh avi oral isolati on incorpora te the same beh avi oral proce s s e s. This hypothesis was not inclu ded in the early ideas abo ut the role of beh avi oral isolati on in spec i a ti on (Dob z h a n s ky 1970), and it has been den i ed by at least one influ en tial evo luti on a ry bi o l ogist (Pa ters on 1989), who argued that within a spec i e s,m a te ch oi ce sel ects for the mean ph en o type ra t h er than an ex treme ph en o type. Previous studies of beh avi oral isolati on bet ween species su gge s t that the factors that cause isolati on evo lved from sex u a lly s el ected ch a racters (e.g., Ri n go 1977, P h elan and Ba ker , Ryan and Rand 1993). One way to test wh et h er sexual sel ecti on forms a con ti n- uum with beh avi oral isolati on is to determine wh et h er females are capable of re s ponding to a hetero s pecific sign a l but nevert h eless prefer a con s pecific sign a l. This test was devel oped and illu s tra ted by Ryan and Rand (1993), wh o s h owed that although female Phys a l a emu s f rogs wo u l d a pproach a playb ack of a hetero s pecific male call wh en it was paired with wh i te noise in a two - s pe a ker te s t,t h ey preferred the call of a con s pecific wh en it was paired with the h etero s pecific call. An o t h er approach is to re a s on that, i f t h e m odels dem on s tra ting that sexual sel ecti on can drive spec i- a ti on are va l i d, it should be po s s i ble to iden tify one or more s ex u a lly sel ected traits within one species (or pop u l a ti on ) that are invo lved also in the failu re of a t tem pts to cross the pop u l a ti on with a close rel a tive (Boa ke et al ). Th i s s econd approach guides my attem pts to understand the ro l e 502 BioScience June 2000 / Vol.50 No. 6

3 of s exual sel ecti on in spec i a ti on in the Hawaiian Dro sop h i - l a. The ideal animal groups for studies of s exual sel ecti on and spec i a ti on would be in the process of d iver ging so that d i f feren ces bet ween them, ra t h er than evo lving after gen e f l ow had disappe a red,m i ght con tri bute to diver gen ce. Fu r- t h erm ore, t h eir mating beh avi or would em ph a s i ze male d i s p l ays ra t h er than a male s abi l i ty to provi de parental care or to pre s ent a high - qu a l i ty terri tory to pro s pective mate s. Su ch species would simplify the stu dy by all owing the re s e a rch er to focus on mating displays. An o t h er import a n t fe a tu re that would en a ble a focus on mating beh avi or is minimal eco l ogical diver gen ce bet ween the two gro u p s being stu d i ed. The list of de s i derata also inclu des the anim a l s amen a bi l i ty to ex peri m en t a ti on and the fe a s i bi l i ty of l a bora tory re a ri n g. The fly species de s c ri bed in this arti cl e m eet many of these cri teri a : Th ey are close rel a tives and the prob a bi l i ty that their diver gen ce was driven by sexual sel ecti on is high, as de s c ri bed in the next secti on. Hawaii and its Dro s oph i l a The rem o teness and geo l ogical history of the Haw a i i a n islands make them an ideal loc a ti on for stu dying evo luti on ( Ca rs on and Clague 1995). The islands form ed over a hot s pot in the eart h s cru s t. The yo u n gest on e, Haw a i i, i s a pprox i m a tely 400,000 ye a rs old; it has active vo l c a n oe s because it is close to the hot spo t. The Pacific plate carri e s the islands to the nort hwe s t, and each island in that directi on is su cce s s ively older; Ka u a i, the oldest high island (on e with mountains), is approx i m a tely 6 mill i on ye a rs old. Th e a to lls and seamounts to the nort hwest of Kauai are the remnants of form er high islands. New islands have becom e ava i l a ble for co l on i z a ti on at intervals of a pprox i m a tely 1 m i ll i on ye a rs. The ori ginal co l onists of the arch i pel a go may h ave come direct ly from a con ti n en t, but many co l on i z a- ti ons have been from older islands within the chain (Wa gner and Funk 1995). In some lineage s, the ance s tor may have a rrived before the pre s ent high islands were form ed. The bi ogeogra phy of Hawaii shows two different patterns of rad i a ti on (Wa gn er and Funk 1995). In some line a ge s, su ch as cri ckets in the genus La u pa l a, co l on i z a ti on of e ach island has re su l ted in ad a ptive rad i a ti on on the island; n ew species have form ed to fill a va ri ety of n i ches (Shaw ). In other lineage s, su ch as the su b group of Dro sop h i - l a that I stu dy, s pec i a ti on is not accom p a n i ed by ad a ptive rad i a ti on that is, the nearest rel a tive of a particular spec i e s is most likely to be found on an ad jacent island in the same kind of h a bitat and using the same kind of host plant ( Ka m bys ellis and Heed 1971, Ka n e s h i ro and Boa ke 1987, De Sa lle 1995). Re s e a rch ers po s tu l a te that each new Haw a i- ian Dro sop h i l a s pecies form ed as a re sult of a single insemi n a ted female being bl own in a storm to a previ o u s ly unocc u p i ed island (Ka n e s h i ro 1976, S p i eth 1981, Ca rs on 1982). The majori ty of re s e a rch on spec i a ti on in Haw a i i a n Dro sop h i l a has been con du cted with the large, p i ctu re - wi n ged flies (Figure 1), p a rti c u l a rly those in the pl a n i ti bi a Fi g u re 2. Hawaiian island chain with the locations of t h e four you n gest spe cies in the p l a n i ti b i a su b grou p. Dro s ophila planiti b i a is thou ght to be ance s tral to the ot h er three spe ci e s, with Dro s ophila heteron eu ra m o s t l i kely having ari sen from Dro s ophila silve s tri s ( De S a ll e and Giddings 1986). su b gro u p. Four spec i e s, Dro sophila differen s, Dro sop h i l a pl a n i ti bi a, Dro sophila silve s tri s, and Dro sophila hetero n eu ra, a re of p a rticular intere s t.e ach species is en demic to a singl e island in the chain (Figure 2), but all four species are similar in their eco l ogy, occ u pying the same alti tu des in cl o u d forest on each island. In all of the spec i e s, feeding and ovi po s i ti on take place on ro t ting stems of plants in the genus C l erm o n ti a ( Ka m bys ellis and Heed 1971, Ca rs on , Ca rs on et al ). Males perch on the bare parts of tree fern fronds and defend terri tories that are cm l ong against other males; females visit and mate with males at these perches (Spieth 1978). Some pairs are unlikely to hybri d i ze in the labora tory for beh avi oral re a s ons (e.g., D. s i lve s tri s and D. d i f feren s; Ka n e s h i ro 1976); in other pairs, hybrid males that do re sult from labora tory mating are sterile (e.g., D. s i lve s tri s and D. pl a n i ti bi a; Crad dock 1974). Th a t t h eir diver gen ce is more beh avi oral than eco l ogical make s these species excell ent for the stu dy of the role of m a ti n g beh avi or in spec i a ti on. Fu rt h erm ore, because the fo u r s pecies repre s ent va rious levels of gen etic and beh avi ora l d iver gen ce (De Sa lle and Giddings 1986), t h ey can be vi ewed as cross secti ons of the process of s pec i a ti on. In trodu ction to D. s i lve s tri s and D. h eteron eu ra The species D. s i lve s tri s and D. h etero n eu ra a re sym p a tric on the island of Hawaii (Figure 2). Ba s ed on gen etic data, t h ey a ppear to be the most cl o s ely rel a ted pair of s pecies in the pl a n i ti bi a su b gro u p. Both species have ch rom o s omal invers i ons that differ from those in most other Haw a i i a n Dro sop h i l a, but all the invers i ons are found in both spec i e s June 2000 / Vol. 50 No. 6 BioScience 503

4 (a) D. s i lve s tri s (b) D. h etero n eu ra Fi g u re 3. Cou rtship seq u en ces for (a) Dro s ophila silve s tri s and (b) Dro s ophila heteron eu ra. In ea ch case, d ep i ctions of t h e m a l e s actions are on the lef t, the fem a l e s on the ri gh t. The thickness of ea ch arrow is propo rtional to the proba bi l i ty of t h e tra n s i ti o n. Beh avi o ral acts are descri bed in the text. H U W, h ea d - u n d er- wi n gs; At tem p. Cop., a t tem pted copu l a ti o n. Da t a from Hoikkala and Wel bergen (1995). Reprodu ced with permission from An n eli Ho i k k a l a, Un ivers i ty of Ou l u, Fi n l a n d. ( Ka n e s h i ro et al ). No unique all oz ymes have been found in ei t h er species (Sene and Ca rs on 1977). The divergen ce bet ween the species measu red with mitoch on d ri a l D NA (mtdna) is approx i m a tely 2%, wh i ch con trasts wi t h a nearly 7% diver gen ce from an out group species (De Sa ll e and Giddings 1986). These species are very similar eco l ogi c a lly and in thei r co u rtship displays. Males of D. s i lve s tri s and D. h etero n eu ra h ave been found displaying in the same patch of woods and even on the same tree fern (Conant 1978). D. s i lve s tri s to l- era tes som ewhat high er alti tu des than D. h etero n eu ra; a ll k n own pop u l a ti ons of D. h etero n eu ra a re sym p a tric with D. s i lve s tri s. The two species differ on ly sligh t ly in the com pon ents of co u rtship beh avi or that are vi s i ble to hu m a n s,t h a t i s,t h ey differ in time bu d gets ra t h er than in the pre s en ce or a b s en ce of p a rticular beh avi or patterns (Wa t s on 1979, Boa ke and Hoikkala 1995, Hoikkala and Wel ber gen 1995). In con trast to co u rtship beh avi or, h owever, male aggre s- s ive beh avi or and its assoc i a ted morph o l ogy diver ge to a s triking ex tent (Spieth 1981). D. s i lve s tri s males rear up on t h eir hind legs while shoving each other, but D. h etero n eu ra males cro u ch and push each other with their wi de head s ( S p i eth 1981). Male and female D. s i lve s tri s h ave ro u n d h e ad s, the norm for mem bers of the genu s. Male D. h et - ero n eu ra h ave stalk eyes that make them appear to have h a m m erh e ads (Figure 1). De s p i te inferen ces that the broad h e ad is invo lved in mate ch oi ce (Spieth 1981), data that su pport that inferen ce have been publ i s h ed on ly recen t ly ( Boa ke et al ). D. s i lve s tri s and D. h etero n eu ra a re capable of produ c i n g com p l etely vi a ble and fertile hybri d s, but mati n gs bet ween females of D. h etero n eu ra and males of D. s i lve s tri s a re ra re. The rec i procal cro s s, wh i ch is easy to produ ce in the labora tory, accounts for all hybrids that have been co ll ected in the field (Ca rs on et al ). This unambiguous beh avi ora l a s ym m etry has been found in many labora tory tests wi t h these species (Crad dock 1974, Val 1977, S p i eth 1978, Pri ce and Boa ke 1995, Fra s er and Boa ke 1997). Thu s, a l t h o u gh the species are morph o l ogi c a lly disti n ct, t h ey show little gen etic diver gen ce, no po s t z ygo tic isolati on (Ah e a rn and Tem p l eton 1989), and incom p l ete beh avi oral isolati on. Cl e a rly, t h ey are sti ll in the process of d iver gi n g. Su ch a pair is ideal for examining how co u rtship signals con tri bute to f a i lu res to hybri d i ze. Inve s ti ga tions of beh avi o ral isol a ti o n One approach to iden ti f ying the traits that con tri bute to beh avi oral isolati on is to con du ct det a i l ed analyses of t h e co u rtship beh avi or of e ach spec i e s, test for differen ce s,a n d i n fer that these differen ces could be invo lved in spec i a ti on. The sequ en ce of co u rtship iden ti f i ed for D. s i lve s tri s and D. h etero n eu ra is shown in F i g u re 3. In both spec i e s, co u rt s h i p i nvo lves an approach in wh i ch the male produ ces a su b- s tra te vi bra ti on call ed a p u rr ( Hoy et al ), and one or both in the pair may wave both wi n gs in large vertical circl e s. The approach is fo ll owed by a facing or circling ph a s e in wh i ch the male faces the fem a l e s head and ei t h er back s up as she approaches or circles to her side while con ti nu i n g 504 BioScience June 2000 / Vol. 50 No. 6

5 to purr. Th en the male circles all the w ay behind the female and takes a h e ad - u n der- wi n gs po s tu re in wh i ch his h e ad is close to the ti p of h er abdom en. If the female does not Ta ble 1. a Sounds produ ced by male Dro s ophila silve s tri s and Dro s ophila heteron eu ra. b P u rr i n g Wing vibration S p e c i e s I n t e rpulse interval (ms) B u rst length (ms) Frequency range (Hz) D. h e t e r o n e u r a 55.0 ± ± D. s i l v e s t r i s 46.2 ± ± a Data from Hoikkala and Welbergen (1995). b Means and standard deviations from at least five males per sample. f ly or run aw ay, the male ex tends both wi n gs latera lly and produ ces a hum by vi bra ting his wi n gs simu l t a n eo u s ly. He t h en ex tends his front legs and brushes the sides of h er a b dom en (Figure 3 does not list this acti on because it is difficult to verify on a vi deo t a pe ), fo ll owing wh i ch he may a t tem pt to cop u l a te. The female plays an active role in advancing co u rtship by producing displays that appear to el i c- it an escalati on of co u rtship from the male; n a m ely, s h e m a kes a lu n ging moti on call ed s l a s h i n g and stands sti ll wh en the male is in the head - u n der- wi n gs po s tu re (Boa ke and Hoikkala 1995, Hoikkala and Wel ber gen 1995). An n eli Hoikkala and her co lleagues at the Un ivers i ty of O u lu in Finland have de s c ri bed the co u rtship of s evera l pl a n i ti bi a- group species in detail and iden ti f i ed differen ce s a m ong the species (Hoikkala and Ka n e s h i ro 1993, Hoi k k a l a et al , Hoikkala and Wel ber gen 1995). Al t h o u gh the gen eral form of co u rtship is similar for D. s i lve s tri s and D. h etero n eu ra, t h ey differ in their airborne and su b s tra te co u rtship sounds and in the time devo ted to the head - u n der- wi n gs po s tu re and to wing vi bra ti on du ri n g co u rtship (Ta ble 1, F i g u re 3). These ob s erva ti ons su gge s t that the head - u n der- wi n gs po s tu re or wing vi bra ti on co u l d d rive reprodu ctive isolati on for this pair. My stra tegy is to examine co u rtship in pairi n gs bet ween s pec i e s, i den tify the stages at wh i ch co u rtship breaks down, and then inve s ti ga te these stages ex peri m en t a lly. Wh en Don Pri ce (now at the Un ivers i ty of Hawaii at Hilo) and I paired D. s i lve s tri s males with D. h etero n eu ra fem a l e s, we fo u n d that co u rtship failed because it did not begi n. Males did not re ach the head - u n der- wi n gs or wi n g - vi bra ti on stage s ; a t m o s t, 10% of a pproaches re su l ted in co u rt s h i p. However, on ce co u rtship had beg u n, it was as likely as the rec i proc a l p a i ring to re sult in cop u l a ti on (Pri ce and Boa ke 1995). Th e s trong role of the approach stage was su rpri s i n g : Because of the differen ces in co u rtship beh avi or bet ween the spec i e s that Hoikkala had ob s erved (Hoikkala and Wel ber gen ), we had anti c i p a ted that wing vi bra ti on would be c ri ti c a l. We propo s e, h owever, that the earliest stage of co u rtship is most cri tical to beh avi oral isolati on, a l t h o u gh l a ter stages may also con tri bute (Pri ce and Boa ke 1995). My stu dents and I are now focusing on the approach s t a ge of co u rt s h i p, wh i ch might invo lve ph erom onal sign a l s, su b s tra te vi bra ti on, and the broad head. Wh en insect co u rtship fails to begi n, an obvious hypothesis is that ph erom ones are invo lved. We are therefore examining the c uticular hyd roc a rbons of e ach species in co ll a bora ti on with a ph erom one bi och em i s t. Di f feren ces bet ween su ch Articles hyd roc a rbons have been implicated in co u rtship isolati on bet ween Dro sophila se ch ell i a and Dro sophila simu l a n s ( Coyne et al ) ; h owever, in D. s i lve s tri s and D. h et - ero n eu ra t h ere is no licking or to u ching of the kind assoc i- a ted with ch emical signaling in other Dro sop h i l a ( S p i et h ). My co lleagues and I are also examining the purr that males produ ce du ring the approach phase of co u rtship (the p u rr differs bet ween the species) (Ta ble 1; Hoy et al , Hoikkala and Wel ber gen 1995). We have com p l eted an analysis of the po s s i ble role for s pecies differen ces in head width in beh avi oral isolati on ( Boa ke et al ). The avera ge male head width is 2.8 mm in D. h etero n eu ra and 1.8 mm in D. s i lve s tri s. We noted that the first vi ew that a female has of an approaching male is his h e ad ; i f she could assess its bre ad t h, she would have a simple cue as to the species of the vi s i tor. To determ i n e wh et h er female D. h etero n eu ra a re capable of assessing head width in con s pecific males, we first te s ted wh et h er head width is sex u a lly sel ected within D. h etero n eu ra. To test for sexual sel ecti on within D. h etero n eu ra, we cons i dered both male male com peti ti on and female mati n g preferen ces (Boa ke et al ). The tests of com peti ti on i nvo lved marking males with a drop of paint on the thora x, t h en pairing them in an ob s erva ti on ch a m ber and watch i n g t h em figh t ; for those pairs that re ach ed the head - s h ovi n g l evel of e s c a l a ti on, we determ i n ed the wi n n er (the one that s tood his ground) and su b s equ en t ly measu red head wi d t h s. In 21 of 25 pairs, the winning male had a broader head,a n d in three cases the head widths were the same. In on ly on e case did the wi n n er have a small er head ; reex a m i n a ti on of this pair reve a l ed that the losing male had received a bi g d rop of paint on his eye ra t h er than his thora x. The tests of female beh avi or invo lved what are call ed n o ch oi ce te s t s, in wh i ch a single pair is placed in an ob s ervati on ch a m ber (Boa ke et al ). These tests are re a s on a bl e for several re a s on s : F i rs t, in the wild the males are so wi dely dispers ed that a female can visit on ly one at a ti m e, a n d s econ d, the ch a m bers are large en o u gh to all ow females to avoid males if t h ey wi s h i n deed, we found no evi den ce of forced cop u l a ti on s, n or have we found all females mati n g. Over the co u rse of 10 days,e ach male was given 10 differen t fem a l e s ; male mating su ccess was scored as the nu m ber of females with wh i ch he mated. At the end of the beh avi ora l te s t s, we measu red the head width and body size of e ach m a l e. We found a high ly significant correl a ti on bet ween male mating su ccess and head width (P = 0.01) that was i n depen dent of body size. Thu s, h e ad width is sex u a lly June 2000 / Vol. 50 No. 6 BioScience 505

6 Fi g u re 4. Gen eti c a lly ba sed va ri a tion in head wi d t h. Fro m l eft to ri gh t,t h e se males are Dro s ophila heteron eu ra, t h e F1 hy brid with D. h eteron eu ra as a mot h er, the F1 hy bri d with Dro s ophila silve s tri s as a mot h er, and D. s i lve s tri s. Ph oto: Todd Ca m pbell. s el ected within D. h etero n eu ra t h ro u gh both interm a l e com peti ti on and female mating preferen ce s. Is it invo lved also in beh avi oral isolati on? To test for a role of h e ad width in beh avi oral isolati on, we a s ked wh et h er female D. h etero n eu ra f a i l ed to mate wi t h male D. s i lve s tri s because their heads were insu f f i c i en t ly broad. Ra t h er than attem pting a su r gical manipulati on of h e ad wi d t h, we took adva n t a ge of the fact that the morph o- l ogical differen ce bet ween the two species shows a large X- ch rom o s omal ef fect (Tem p l eton 1977, Val 1977). Rec i proc a l hybrid males have iden tical auto s omes but differ in the origins of t h eir X ch rom o s om e s, as well as in head width (Figu re 4). Hybrid males with a D. h etero n eu ra m o t h er have broader heads (2.5 mm) than those with a D. s i lve s tri s m o t h er (1.98 mm). Thu s,i f h e ad width is the re a s on that D. h etero n eu ra females failed to mate with D. s i lve s tri s m a l e s, the two hybrids should have lower mating su ccess than D. h etero n eu ra m a l e s, with mating su ccess among the hybri d s d i f fering according to their head wi d t h. However, we fo u n d no differen ces in mating su ccess wh en we te s ted more than 70 males of e ach of the three gen o types (Figure 5; Boa ke et a l ). Nor did we find any evi den ce for X-ch rom o s om a l i n f lu en ces on the stru ctu re of co u rtship beh avi or, wh i ch could have been a com p l i c a ting factor (Boa ke et al ). These re sults for head width are incon s i s tent with the gen eral hypothesis that sexual sel ecti on and beh avi ora l i s o l a ti on form a con ti nu u m. Al t h o u gh it is po s s i ble that the hypothesis is incorrect, it would be nece s s a ry to rej ect the hypothesis for all rel evant sexual signals before it could be dec i s ively rej ected for this species pair. We are s ti ll te s ting po s s i ble roles for ph erom ones and su b s tra te vi bra ti on. It is also po s s i ble that hybrids in the labora tory perform mu ch bet ter than they would in the natu ral habit a t, as has been found for sti ck l eb ack fishes (Ha tf i eld and S ch luter 1999). If that propo s i ti on is tru e, the lack of a d i f feren ce in mating su ccess bet ween hybrid and D. h et - ero n eu ra males might be an arti f act, but the similari ty of m a ting su ccess bet ween the hybrid males would be difficult to ign ore because the rec i procal hybrids are auto s o- m a lly the same. If the gen eral hypothesis is rej ected for D. s i lve s tri s and D. h etero n eu ra, the next logical step wo u l d be to find out wh et h er it is untrue for other pairs of s pec i e s. It would be va lu a ble to learn wh et h er there are a ny taxon omic or eco l ogical factors that would all ow us to pred i ct wh en the hypothesis is likely to be appropri a te. Because we are sti ll te s ting the other cues that may be i nvo lved in the approach that initi a tes co u rt s h i p, we cannot con clu de that the hypothesis is wron g. Perhaps several cues, opera ting in a hiera rchical fashion, con tri bute to beh avi ora l i s o l a ti on (and po s s i bly also to sexual sel ecti on ). Two hypotheses need to be inve s ti ga ted : F i rs t, i f o t h er cues are adequ a te, h e ad width becomes an important cue, as is seen within D. h etero n eu ra; and secon d, o t h er cues may be not m erely adequ a te but capable of overriding the nega tive ef fects of a ro u n d - h e aded male, wh i ch could explain our re sults for the F1 hybrid males. We are also ob s erving po s t a pproach co u rtship in det a i l because it may play a ro l e, even if on ly a secon d a ry on e, i n beh avi oral isolati on. An obvious candidate for furt h er i nve s ti ga ti on is the sound produ ced du ring wing vi bra ti on, wh i ch differs bet ween species (Ta ble 1). However, wh en we ex a m i n ed the intra s pecific mating su ccess of males wh o s e wi n gs had been cl i pped of f,t hus disabling their sound produ cti on, we found no differen ces bet ween cl i pped males and uncl i pped con trols (Boa ke and Po u l s en 1997). Si m i l a r- ly, in inters pecific pairs, cl i pped male D. h etero n eu ra a re as su ccessful as normal males in co u rting D. s i lve s tri s fem a l e s ( Ch ri s tine R. B. Boa ke, u n p u bl i s h ed data). Re aders may have recogn i zed an unsolved qu e s ti on that is clear even from the de s c ri pti ons at the start of this report. If these species have su ch weak beh avi oral barri ers and no po s t z ygo tic isolati on, why are they disti n ct? The natu ra l hybrids that were found were no more than 2% of a ll indivi duals co ll ected (Ca rs on et al ). One hypothesis is that eco l ogical disti n cti ons are stron ger than we think. For ex a m p l e, S p i eth (1981) noted that D. h etero n eu ra ten ded to be found on perches with bri gh ter light than those on wh i ch D. s i lve s tri s were fo u n d. It is also po s s i ble that the a l ti tudinal ra n ge of D. h etero n eu ra u s ed to ex tend lower than that of D. s i lve s tri s, but because the native wet fore s t s at lower alti tu des have been cl e a red for agri c u l tu re, t h i s hypothesis is unte s t a bl e. In the mid-1990s I attem pted to m e a su re the ligh ti n g, tem pera tu re, and veget a ti on den s i ty at male perch site s, but the flies had nearly disappe a red. Th ey are prey for introdu ced yell ow jackets (Ve s pula pen s yl - va n i c a; Foo te and Ca rs on 1995) and their host plants are con su m ed by feral pigs. Thu s, these species are an ex a m p l e of an incre a s i n gly com m on probl em in modern bi o l ogy: the loss of f u n d a m ental inform a ti on abo ut cri tical issu e s because of ex ti n cti on (Grant 2000). Co n cl u s i o n s I have de s c ri bed a stra tegy for iden ti f ying species differen ces in co u rtship and devel oping ex peri m ents to eva lu a te 506 BioScience June 2000 / Vol.50 No. 6

7 Mating success D. heteroneura HS hybrid SH hybrid Male genotype t h eir import a n ce to beh avi oral isolati on. Traits that are s h own to influ en ce beh avi oral isolati on would then be te s t- ed to discover their role in sexual sel ecti on. Su ch stu d i e s cannot be used to determine wh et h er sexual sel ecti on drove s pec i a ti on in the past, but they can su pport or disprove the hypothesis that sexual sel ecti on and beh avi oral isolati on are i n f lu encing the same traits now. Thu s, we can test wh et h er s exual sel ecti on and beh avi oral isolati on are parts of a conti nu u m. For this re s e a rch, it is essen tial to stu dy pairs of s pecies or pop u l a ti ons that are in the process of d iver gi n g ; o t h erwi s e, the ob s erved differen ces could have evo lved after the com p l eti on of s pec i a ti on and not be direct ly invo lved in s pec i a ti on. It is also important to re a l i ze that even if t wo s pecies show a high degree of eco l ogical similari ty, as do the su bj ects of my re s e a rch, the po s s i ble role of eco l ogy in spec i a ti on must not be ign ored. A caveat to this re s e a rch is that it is de s c ri bed as though the females discri m i n a te among po ten tial mate s, wh i ch is com m on ly the case in sexual sel ecti on. However, we know that in these fly spec i e s, male dec i s i ons to adva n ce the progress of co u rtship are influ en ced by signals that they receive from females (Boa ke and Hoikkala 1995, Hoi k k a l a and Wel ber gen 1995). Males of e ach species wi ll approach a nything that is small, d a rk, and moving (including ju m p- ing spiders, with an unfavora ble outcome for the fly ), but t h ey do not re ad i ly adva n ce to the head - u n der- wi n gs stage even with a con s pecific fem a l e. Con s equ en t ly, the ph en omen on that is call ed beh avi oral isolati on is likely to invo lve bi d i recti onal sign a l i n g, with males needing appropri a te signals from females as mu ch as females need correct sign a l s f rom males. An adva n t a ge of vi ewing beh avi oral isolati on and sexual sel ecti on as parts of a con ti nuum is that we no l on ger need to con s i der beh avi oral isolati on as a uniqu e tra i t ; ra t h er, it is a mismatch bet ween the sexual signals that a re used by each spec i e s. Our goal is to iden tify the mism a tch. The re s e a rch de s c ri bed above is part of a larger inve s ti gati on of the process of s pec i a ti on. Recent gen etic model s h ave foc u s ed on the role that sexual sel ecti on could play in Fi g u re 5. Ma ting su ccess of males of t h ree gen otypes wi t h f emale Dro s ophila heteron eu ra (data from Boa ke et al ). From left to ri gh t, the data are from D. h eteron eu ra, the F1 hy brid with D. h eteron eu ra as a mot h er (HS), and the F1 hy brid with Dro s ophila silve s tri s as a mot h er (SH); differen ce s a re not significant at P = The mean head widths from lef t to ri ght are 2.81 mm, 2.5 mm, and 1.98 mm. The sample size s from left to ri ght are 78, 7 9, and 75 males. s pec i a ti on (Tu rn er and Bu rrows 1995), while others have ad d re s s ed con ten tious issues abo ut the nu m ber of gen e s that could influ en ce spec i a ti on (Gavri l ets and Ha s ti n gs ). These mathem a tical models have clear and te s t a bl e a s su m pti ons and pred i cti on s. The most appropri a te tra i t s with wh i ch to test su ch models are those that have been dem on s tra ted to play a cri tical role in an on going process of s pec i a ti on, using met h ods su ch as the ones de s c ri bed in this a rti cl e. Ack n owl e d gm en t s This re s e a rch has ben ef i ted gre a t ly from the input from my co ll a bora tors, as well as the en co u ra gem ent and advi ce of Hamp Ca rs on and Ken Ka n e s h i ro, who introdu ced me to the flies and whose work sti mu l a ted many of my qu e s ti on s. Nu m erous under gradu a te stu dents have parti c i p a ted in the e s s en tial but tedious work of s tock care. Jessica Bi er, Rebecca Ch a s a n, Susan Fo s ter, All en Moore, and several anonymous revi ewers provi ded va lu a ble su gge s ti ons rega rd i n g the tex t. While wri ti n g, I was assisted by funding from the US Na ti onal Scien ce Fo u n d a ti on (grant no. I B N ). Ref eren ces ci te d Ah e a rn JN, Tem p l eton A R In ters pecific hybrids of Dro sophila het - ero n eu ra and D. s i lve s tri s. I. Co u rtship su cce s s. Evo luti on 43: Boa ke CRB, Hoikkala A Co u rtship beh aviour and mating su ccess of wi l d - c a u ght Dro sophila silve s tri s m a l e s. Animal Beh aviour 49: Boa ke CRB, Po u l s en T Correl a tes versus pred i ctors of co u rt s h i p su cce s s : Co u rtship song in Dro sophila silve s tri s and D. h etero n eu ra. Animal Beh aviour 54: Boa ke CRB, De An gelis M, An d re adis DK Is sexual sel ecti on and s pecies recogn i ti on a con ti nuum? Ma ting beh avi or of the stalk-eyed f ly Dro sophila hetero n eu ra. Proceed i n gs of the Na ti onal Ac ademy of S c i en ces of the Un i ted States of Am erica 94: Boa ke CRB, Pri ce DK, An d re adis DK In h eri t a n ce of beh avi o u ra l d i f feren ces bet ween two interferti l e,s ym p a tric spec i e s, Dro sophila sil - ve s tri s and D. h etero n eu ra. Hered i ty 80: Ca rs on HL Evo luti on of Dro sop h i l a on the newer Hawaiian vo l c a- n oe s. Hered i ty 48: Ca rs on HL, Clague DA G eo l ogy and bi ogeogra phy of the Haw a i i a n i s l a n d s. Pa ges in Wa gn er W L, Funk VA, ed s. Hawaiian Bi ogeogra phy: Evo luti on on a Hot Spot Arch i pel a go. Wa s h i n g ton (DC ) : Sm i t h s onian In s ti tuti on Pre s s. Ca rs on HL, Ka n e s h i ro KY, Val FC Na tu ral hybri d i z a ti on bet ween the sym p a tric Hawaiian species Dro sophila silve s tri s and Dro sop h i l a h etero n eu ra. Evo luti on 43: Cl a rke BC, Grant PR, ed s Evo luti on on islands. P h i l o s oph i c a l Tra n s acti ons of the Royal Soc i ety of Lon don B Bi o l ogical Scien ce s : Conant P Lek beh avi or and eco l ogy of t wo sym p a tric hom o s e- June 2000 / Vol. 50 No. 6 BioScience 507

8 qu en tial Hawaiian Dro sop h i l a: Dro sophila hetero n eu ra and Dro sop h i l a s i lve s tri s. Ma s ter s thesis. Ma n oa (HI): Un ivers i ty of Haw a i i. Coyne JA, Cri t ten den A P, Mah K G en etics of a ph erom onal differen ce con tri buting to reprodu ctive isolati on in Dro sop h i l a. S c i en ce 265: Crad dock EM Reprodu ctive rel a ti onships bet ween hom o s equ en ti a l s pecies of Hawaiian Dro sop h i l a. Evo luti on 28: D a rwin C The De s cent of Man and Sel ecti on in Rel a ti on to Sex. Lon don : Mu rray. De Sa lle R Mo l ecular approaches to bi ogeogra phic analysis of Hawaiian Dro s oph i l i d ae. Pa ges in Wa gn er W L, Funk VA, ed s. Hawaiian Bi ogeogra phy: Evo luti on on a Hot Spot Arch i pel a go. Wa s h- i n g ton (DC ) : Sm i t h s onian In s ti tuti on Pre s s. De Sa lle R, G i d d i n gs LV Di s cord a n ce of nu clear and mitoch on d ri a l D NA phyl ogenies in Hawaiian Dro sop h i l a. Proceed i n gs of the Na ti onal Ac ademy of S c i en ces of the Un i ted States of Am erica 83: Dob z h a n s ky T G en etics of the Evo luti on a ry Proce s s. New York : Co lu m bia Un ivers i ty Pre s s. F i s h er RA The Gen etical Th eory of Na tu ral Sel ecti on. 2nd ed. New York : Dover. Foo te D, Ca rs on HL Dro sop h i l a as mon i tors of ch a n ge in Haw a i i a n eco s ys tem s. Pa ges in LaRoe ET, Fa rris GS, Pu ckett CE, Dora n P D, Mac MJ, ed s. Our Living Re s o u rce s : A Report to the Na ti on on the Di s tri buti on, Abu n d a n ce, and Health of U. S. P l a n t s, An i m a l s, a n d E co s ys tem s. Wa s h i n g ton (DC ) : US Dep a rtm ent of the In teri or, Na ti onal Bi o l ogical Servi ce. Fo s ter SA, S cott RJ, Cre s ko WA Ne s ted bi o l ogical va ri a ti on and spec i a ti on.p h i l o s ophical Tra n s acti ons of the Royal Soc i ety of Lon don B Bi o l ogical Scien ces 353: Fra s er I, Boa ke CRB Beh avi oral isolati on, test de s i gn s, and Ka n e s h i- ro s hypo t h e s i s. Am erican Na tu ralist 149: Futuyma DJ, Mayer GC Non - a ll op a tric spec i a ti on in animals. System a tic Zoo l ogy 29: G avri l ets S, Ha s ti n gs A Fo u n der ef fect spec i a ti on : A theoreti c a l re a s s e s s m en t. Am erican Na tu ralist 147: Grant PR What does it mean to be a natu ralist at the end of t h e t wen ti eth cen tu ry? Am erican Na tu ralist 155: Grant PR, Grant BR S pec i a ti on and hybri d i z a ti on in island bi rd s. P h i l o s ophical Tra n s acti ons of the Royal Soc i ety of Lon don B Bi o l ogical Scien ces 351: Ha rri s on RG Linking evo luti on a ry pattern and proce s s : The rel e- va n ce of s pecies con cepts for the stu dy of s pec i a ti on. Pa ges in How a rd DJ,Berl och er SH, ed s.e n dless Form s :S pecies and Spec i a ti on. Ox ford : Ox ford Un ivers i ty Pre s s. Ha tf i eld T, S ch luter D E co l ogical spec i a ti on in sti ck l eb ack s : E nviron m en t - depen dent hybrid fitn e s s. Evo luti on 53: Hoikkala A, Ka n e s h i ro KY Ch a n ge in the sign a l - re s ponse sequ en ce re s pon s i ble for asym m etric isolati on bet ween Dro sophila pl a n i ti bi a and Dro sophila silve s tri s. Proceed i n gs of the Na ti onal Ac ademy of S c i- en ces of the Un i ted States of Am erica 90: Hoikkala A, Wel ber gen P Si gnals and re s ponses of females and males in su ccessful and unsu ccessful co u rtships of t h ree Hawaiian lekm a ting Dro sop h i l a s pec i e s. Animal Beh aviour 50: Hoikkala A, Ka n e s h i ro KY, Hoy RR Co u rtship son gs of the pictu re - wi n ged Dro sophila pl a n i ti bi a su b group spec i e s. Animal Beh aviour 47: Hoy RR, Hoikkala A, Ka n e s h i ro KY Hawaiian co u rtship son gs : Evoluti on a ry innova ti on in com mu n i c a ti on signals of Dro sop h i l a. S c i en ce : Ka m bys ellis MP, Heed W B S tudies of oogenesis in natu ral pop u l a- ti ons of D ro s oph i l i d ae. I. Rel a ti on of ova rian devel opm ent and ecol ogical habitats of the Hawaiian spec i e s. Am erican Na tu ralist 105: Ka n e s h i ro KY Et h o l ogical isolati on and phyl ogeny in the pl a n i ti bi a su b group of Hawaiian Dro sop h i l a. Evo luti on 30: Sexual isolati on,s pec i a ti on, and the directi on of evo luti on. Evo luti on 30: Ka n e s h i ro KY, Boa ke CRB Sexual sel ecti on and spec i a ti on : Is su e s ra i s ed by Hawaiian Dro sop h i l a. Trends in Eco l ogy and Evo luti on 2: Ka n e s h i ro KY, G i llespie RG, Ca rs on HL Ch rom o s omes and male genitalia of Hawaiian Dro sop h i l a: Tools for interpreting phyl ogeny and geogra phy. Pa ges in Wa gn er W L, Funk VA, ed s. Haw a i i a n Bi ogeogra phy: Evo luti on on a Hot Spot Arch i pel a go. Wa s h i n g ton ( DC ) : Sm i t h s onian In s ti tuti on Pre s s. L a n de R Models of s pec i a ti on by sexual sel ecti on on po lygen i c tra i t s. Proceed i n gs of the Na ti onal Ac ademy of S c i en ces of the Un i ted S t a tes of Am erica 78: L a n de R, Ki rk p a tri ck M E co l ogical spec i a ti on by sexual sel ecti on. Jo u rnal of Th eoretical Bi o l ogy 133: Mayr E Animal Species and Evo luti on. Ca m bri d ge (MA): Ha rva rd Un ivers i ty Pre s s. Pa lu m bi SR S pecies form a ti on and the evo luti on of ga m ete recogn i ti on loc i. Pa ges in How a rd DJ, Berl och er SH, ed s.e n dl e s s Form s :S pecies and Spec i a ti on. Ox ford : Ox ford Un ivers i ty Pre s s. Pa ters on HEH A vi ew of s pec i e s. Pa ges in Goodwin B, Sibatani A, Web s ter G, ed s. Dynamic Stru ctu res in Bi o l ogy. E d i n bu r gh ( U K ) :E d i n bu r gh Un ivers i ty Pre s s. P h elan PL, Ba ker TC Evo luti on of male ph erom ones in moths: Reprodu ctive isolati on thro u gh sexual sel ecti on? Scien ce 235: Pri ce DK, Boa ke CRB Beh avi oral reprodu ctive isolati on in Dro sophila silve s tri s, D. h etero n eu ra, and their F1 hybrids (Di ptera : D ro s oph i l i d ae ). Jo u rnal of In s ect Beh avi or 8: Ri n go JM Why 300 species of Hawaiian Dro sop h i l a? The sex u a l s el ecti on hypo t h e s i s. Evo luti on 31: Ryan MJ, Rand A S S pecies recogn i ti on and sexual sel ecti on : A unit a ry probl em in animal com mu n i c a ti on. Evo luti on 47: S ch luter D E co l ogical causes of s pec i a ti on. Pa ges in How a rd DJ,Berl och er SH, ed s.e n dless Form s :S pecies and Spec i a ti on. Ox ford : Ox ford Un ivers i ty Pre s s. Sene FM, Ca rs on HL G en etic va ri a ti on in Hawaiian Dro sop h i l a. I V. All oz ymic similari ty bet ween D. s i lve s tri s and D. h etero n eu ra f rom the Island of Haw a i i.g en etics 86: S h aw KL Bi ogeogra phic patterns of t wo indepen dent Haw a i i a n c ri cket rad i a ti ons (La u pa l a and Pro gn a t h o gryll u s). Pa ges in Wa gn er W L, Funk VA, ed s. Hawaiian Bi ogeogra phy: Evo luti on on a Hot Spot Arch i pel a go. Wa s h i n g ton (DC ) : Sm i t h s onian In s ti tuti on Pre s s. S p i eth HT Co u rtship patterns and evo luti on of the Dro sophila adi - a s tol a and pl a n i ti bi a s pecies su b gro u p s. Evo luti on 32: Dro sophila hetero n eu ra and Dro sophila silve s tri s: He ad s h a pe s, beh avi or and evo luti on. Evo luti on 35: Tem p l eton A R An a lysis of h e ad shape differen ces bet ween two i n terfertile species of Hawaiian Dro sop h i l a. Evo luti on 31: Tu rn er GF, Bu rrows MT A model of s ym p a tric spec i a ti on by sex u- al sel ecti on. Proceed i n gs of the Royal Soc i ety of Lon don B Bi o l ogi c a l S c i en ces 260: Val FC G en etic analysis of the morph o l ogical differen ces bet ween t wo infertile species of Hawaiian Dro sop h i l a. Evo luti on 31: Wa gn er W L, Funk VA, ed s Hawaiian Bi ogeogra phy: Evo luti on on a Hot Spot Arch i pel a go. Wa s h i n g ton (DC ) : Sm i t h s onian In s ti tuti on Pre s s. Wa t s on GF On prem a ting isolati on bet ween two cl o s ely rel a ted s pecies of Hawaiian Dro sop h i l a. Evo luti on 33: We s t - E berh a rd MJ Sexual sel ecti on,s ocial com peti ti on, and spec i- a ti on.q u a rterly Revi ew of Bi o l ogy 58: BioScience June 2000 / Vol. 50 No. 6