The Condor 101:518-528 0 The Cooper Ornithological Society 1999

THE FREQUENCY AND FUNCTION OF AQUATIC COURTSHIP AND COPULATION IN LEAST, CRESTED, WHISKERED, AND PARAKEET AUKLETS ’

E M. HUNTER Department of Zoology, Downing Street, Universityof Cambridge, Cambridge, CB2 3EJ, UK, e-mail: [email protected]

I. L. JONES Department of Biology, Memorial Universityof Newfoundland,St. Johns,Newfoundland, AIB 3X7, Canada

Abstract. Least ( pusilla), Crested (A. cristatella), Whiskered (A. pygmaea), and ParakeetAuklets (Cyclorrhynchuspsittaculu) all engagein aquaticcourtship and copulation behavior. We quantified auklet sexual behavior at sea to obtain comparative information related to sexual selection and to evaluate why auklets choose this unusual location for mating. Auklet courtshipinvolved a variety of stereotypeddisplays. Although similar court- ship displays occurredboth at the colony on land and at sea, copulationtook place only on the sea. Courtship and copulation was frequently disrupted by extrapair males. Extrapair copulation attempts occurred in all species and apparently successfulunforced extrapair copulationswere observed in Least, Crested, and Whiskered Auklets, suggestingthat sperm competition is a feature of their mating systems. Male risk water damage to their sperm if they mount their female partners at sea and cause their cloacas to become sub- merged during insemination. Male auklets, lacking an intromittent organ, achieved cloaca1 contact without mounting the female so that their cloacas remained unsubmerged.Males positioned themselvesbehind their partnersand rapidly flapped their wings, creating lift to bring their cloacas up against their partner’s cloacas,pushing upwards and ensuringinsem- ination occurredaway from the seasurface. Possible hypotheses to accountfor exclusively aquaticcopulation include and sexualharassment avoidance, and femaletesting of males,but a definitiveexplanation for this phenomenonremains elusive. Key words: Alcidae, auklet, courtship,copulation, sexual selection,sperm competition.

INTRODUCTION detailed observation of pairs at sea as well as on land is essential. Darwin (1871) suggested that sexual selection Observations of sexual behavior at sea also would give rise to variation in both ornamental would reveal how speciesthat copulate on water traits and courtship behaviors among closely re- overcome the resulting problem of sperm trans- lated species. In the majority of species fer. When a male mounts a female on the in which courtship and copulation have been de- water, her cloaca is likely to become submerged scribed, copulation and associatedbehavior oc- (Bums et al. 1980). If sperm are inseminated by cur at or near the breeding site on land (Hatch- cloaca1contact, the usual avian method, they are well 1988, Wagner 1991, Hunter et al. 1992). likely to be washed away, diluted, or damaged For that copulate at sea, very little is by the surrounding water, resulting in reduced known about patterns of courtship and the be- likelihood of fertilization. Male waterfowl (An- haviors involved in copulation, due to the diffi- seriformes), which copulate on water, have an culties of observing complex behavior at sea. In intromittent organ which allows direct place- a study of Black-legged Kittiwake (Rissa tridac- ment of sperm into the female’s reproductive @la) behavior in offshore flocks, Daniels et al. tract (McKinney et al. 1983). Lake (1981) sug- (1994) found that courtship displays normally gested that internal sperm transfer by way of an seen on land were modified at sea and that one intromittent organ overcomes the problem of previously unrecorded courtship display oc- water damage. In a review of birds that regularly curred exclusively on the water. This suggests spend periods of time on water during the breed- that to fully evaluate seabird courtship behavior, ing period, Briskie and Montgomerie (1997) found that bird speciesthat have an intromittent organ tended to copulate on water, whereas ’ ’ Received7 July 1998.Accepted 19 January1999. those that lack an intromittent organ copulate on AUKLET COURTSHIP AND COPULATION 519 land. This supported the possibility that the in- METHODS tromittent organ is an important adaptation for We made observations of all species of auklets birds that copulate on or in water. However, at their ocean staging area adjacent to Main Ta- Briskie and Montgomerie (1997) found a num- lus on Buldir Island, Alaska (52”2’N, 17Y5’E). ber of important exceptions to the water damage Observations were made from a hide situated 20 hypothesis, in particular, phalaropes, pelicans, m above sea level, directly overlooking the - and alcids, species that lack intromittent organs let staging flock. Least and Crested Auklets but copulate on water. formed a mixed-species aggregation each mom- Among the alcids, eight species copulate at ing during the breeding season which main- sea (Gaston and Jones 1998), including four spe- tained a distance of between 100 m and 400 m cies of auklets (Least Aethia pusillu, Crested A. offshore. Parakeet Auklets formed several small- cristatella, Whiskered A. pygmaea, and Parakeet er single species flocks closer inshore (5 to 50 Auklet Cyclorrhynchuspsittuculu), three species m offshore), as did Whiskered Auklets at an in- of puffins (Tufted Fratercula cirrhata, Horned termediate distance (30 to 200 m offshore). All F. corniculata, and F. arctica), behaviors were clearly visible at these distances and the Marbled Murrelet (Brachyramphusmar- using either a 30X by 80 mm Kowa TSN-4 or morutus). Auklets (Alcidae, tribe Aethiini) are 20X by 60 mm Kowa TSN-2 telescope. small socially monogamous seabirds of the North Pacific, Bering and Okhotsk Seas (Gaston We restricted our observations to days with and Jones 1998). Least and Crested Auklets are calm sea conditions and good visibility, to allow diurnal and highly colonial. These two species clear viewing of focal pairs. We made observa- occur together at many colonies, but Crested tions on 14 days between 14 May and 17 June Auklets are larger, more sexually dimorphic and 1995, on 6 days between 5 and 13 June 1996, aggressive than Least Auklets. Whiskered Auk- and on 12 days between 12 May and 10 June lets are mainly nocturnal on land, and nest in 1997, during the period of peak pre-laying court- both dense and dispersedcolonies, whereas Par- ship behavior for auklet species at Buldir. E M. akeet Auklets are diurnal but nest in small dis- Hunter made additional observations of Least persed colonies. These four species show vary- Auklets and Parakeet Auklets on the staging ing levels of ornamentation. Whiskered Auklets area offshore from the Antone Lake auklet col- are highly ornamented, with a spectacular fore- ony on St. Paul Island, Pribilof Islands, Alaska head crest and three pairs of long white facial (57”08’N, 170”17’W) on 12 days during the pe- plumes; Crested Auklets have a large crest, one riod 6 to 20 May 1996, from a vehicle posi- pair of white facial plumes and a bright orange tioned on the Antone Lake sea wall 15 m above bill; Least Auklets have one pair of small white sea level, overlooking the staging flock. facial plumes and a red bill with a knob-like We observed the birds for 4 to 5 hr each day ornament; whereas Parakeet Auklets have a red during their morning activity period, usually bill and a single pair of white facial plumes. Var- starting soon after dawn when there was suffi- ious studieshave described courtship displays of cient light for viewing. Focal pairs were chosen Least and Crested Auklets that occur on land at random and observed for up to 5 min. Pairs (Jones and Montgomerie 1992, Jones and Hunt- that flew from the staging area or were lost er 1993). However, courtship and copulation be- among other birds in the flock before 2 min of havior at sea has not been quantified for any the observation period had elapsed were not in- auklet species and more specifically it is not cluded in the analysis. A pair was identified as known how male auklets avoid water damage two birds maintaining close proximity, within a during mating at sea. maximum distance of two to three bird-lengths, The aims of our study were to quantify and usually closer, and characteristically well sepa- compare aquatic courtship and copulation be- rated from other auklets at the staging area. We havior of Least, Crested, Whiskered, and Para- could not usually identify these couples as mat- keet Auklets, and to addresstwo specific ques- ed pairs, breeding together at the colony. How- tions: (1) how do male auklets, which lack an ever, we regularly observed mated pairs, some intromittent organ, avoid water damage to their of which were identifiable individuals from our sperm? and (2) why do auklets copulate exclu- color-marked populations at Main Talus, depart- sively at sea? ing from our study plot and flying together to 520 F. M. HUNTER AND I. L. JONES the staging area where they landed on the sea RESULTS and engaged in identical behavior to our focal COURTSHIP DISPLAYS pairs. We assumed that conspecific individuals that disrupted our focal pairs represented extra- We described and quantified the behavior of 404 pair individuals. Least, 246 Crested, 43 Whiskered, and 224 Par- We treated each focal pair as an independent akeet Auklet pairs at sea. We identified six dis- data point because we believe that the probabil- tinct courtship displays performed by auklets on ity of observing the same pair more than once the sea: hunch, pursuit, ruff-sniff, neck-twist, during our study was very low. This belief is head-bob, and mutual-court (Fig. la-d). The based on the large auklet populations at our hunch, pursuit, and mutual-court displays were study sites (>200,000 pairs at Buldir, >25,000 used by all four species,although pursuit behav- pairs at St. Paul Island), estimates of which are ior was most prevalent in Crested Auklets. The based on net movement counts of Least and neck-twist display was performed almost exclu- Crested Auklets by Byrd et al. (1983) and ob- sively by Crested Auklets, the ruff-sniff display servations by us, and on the large number of was unique to Crested Auklets, and the head- pairs that were present in the staging flocks on bob was performed only by Whiskered Auklets any day. Estimated numbers of birds in the Main (Fig. 2a-e). The hunch display was similar to Talus staging flock were 21,000 one performed during Least and pairs, 43,000 Crested Auklet pairs, 5,000 Whis- courtship on land (Jones 1993a, 1993b). One or both pair members leaned forward toward the kered Auklet pairs, and 500 Parakeet Auklet other, with necks stretched outward and held pairs, whereas estimated numbers in the Antone close to the water, bills oriented upwards, and Lake staging flock were 2,000 Least Auklet nape and upper back feathers erected (Fig. la). pairs and 500 Parakeet Auklet pairs. Hunch displays often involved one pair member We recorded all courtship displays and both pursuing the other with neck down and head up attempted and successfulpair and extrapair cop- towards its partner. We referred to this combi- ulations (EPCs) that occurred during the obser- nation of hunched posture and close following vation period of each focal pair. Stereotyped as the pursuit display. All four speciesof auklets courtship displays were identified early in the engaged in both hunch and pursuit displays study allowing both of us to record comparable (Figs. 2a, b). Hunch displays were most preva- information on display frequency. It was not lent in Least Auklets with 42.1% (170/404) of possible for us to identify whether sperm was focal pairs engaging in this behavior. Pursuit dis- transferred during cloaca1 contact nor indeed plays were most frequent in Crested Auklets whether cloaca1contact was accurately achieved with 73.6% (181/246) of focal pairs displaying in every case. The best measure of copulation this behavior. In Whiskered Auklets, the successthat could be attained was apparent clo- hunched posture occurred almost exclusively as acal contact or behaviorally successful copula- part of pursuit displays (hunch 2.3%, l/43; pur- tion. suit 20.9%, 9/43). During observations of auklet behavior on An exaggerated form of close pursuit behav- land, we looked for evidence of copulation and ior was observed in Crested Auklets at sea. One attempted copulations throughout the 1990- individual would closely follow, with its body 1998 breeding seasons. To evaluate the possi- alongside but partly behind, its partner, its breast bility of copulation within nesting crevices, we resting on the partner’s back and its bill nestling measured the greatest vertical height from ceil- into the neck feathers of its partner (Fig. lb). ing to floor of the nesting chambers of 38 Least, This latter behavior appears to be a modified 32 Crested, 54 Whiskered, and 11 Parakeet Auk- form of the ruff-sniff courtship display which let crevices using a piece of string with a weight Crested Auklets perform during land-based attached. The length of the string was then mea- courtship (Zubakin 1990, Jones 1993a). Both sured to the nearest 1 mm using a ruler. The male and female Crested Auklets took the active sample was limited to crevices that had entranc- “nuzzling” role in ruff-sniffing behavior at sea. es large enough to allow the passage of our In Crested Auklets, 8.8% (16081) of pursuits hands and with nesting chambers within an involved ruff-sniffing. arm’s length of the entrance. Another courtship behavior almost entirely AUKLET COURTSHIP AND COPULATION 521

FIGURE 1. Courtship and copulation behavior in auklets. Courtship displays illustrated for the species that engages in that display most frequently. (a) Hunch display in Least Auklets. (b) Pursuit with ruff-sniff behavior in Crested Auklets. (c) Neck-twist display in Crested Auklets. (d) Mutual-court display in ParakeetAuklets. (e) Behaviorally successfulcopulation involving apparentcloaca1 contact in Least Auklets. limited to Crested Auklets was the neck-twist 1993a). A small number of Least Auklets en- display, in which the male and female mutually gaged in this display at sea (2.0%, S/404 of intertwined their necks, often while facing each Least Auklet focal pairs, Fig. 2c) but we never other and circling each other on the water (Figs. observed it in Whiskered or Parakeet Auklets. lc, 2~). This display also is performed frequent- The head-bob display was unique to Whiskered ly by Crested Auklets courting on land (Jones Auklets (Fig. 2d; V. Zubakin, unpubl. data). 522 F. M. HUNTER AND I. L. JONES

b

1001 C d 80 -I ;; 2 60.

Y Q) 40- z @ 20-

1001 e f

r

Le Au CrAu WhAu PaAu Le Au CrAuI- Wh Au IlPaAu Species Species

FIGURE 2. Proportion of pairs of Least (LeAu), Crested (CrAu), Whiskered (WhAu), and Parakeet (PaAu) Auklets that exhibited (a) hunch display, (b) pursuit behavior, (c) neck-twist display, (d) head-bob display, (e) mutual-court display, and (f) any courtshipdisplay.

While facing its partner, one or both pair mem- the form of both members of the pair facing and bers rhythmically and repeatedly raised and low- vocalizing towards each other (Fig. Id). This ered the head, often with the head stretchedfor- display was recorded in all four speciesof auklet ward. Of all focal pairs of Whiskered Auklets although it was particularly prevalent in Para- observed, 67.4% (29143) were seen to engage in keet Auklets with 63.7% (121/190) of observed a head-bobbing display. Mutual-courting took pans on Buldir and 38.2% (13134) of pairs on AUKLET COURTSHIP AND COPULATION 523

TABLE 1. Heights of mounted pairs of auklets in relationto spaceavailable in the nestingcrevice.

Estimated w crev,ces height of Crevice height (mm) smaller than copulating height of Species pair (mm)” Meall SD Range n copulatmg pair Least Auklet 120 99.8 37.8 44-192 38 73.7 Crested Auklet 170 153.4 52.3 98-329 32 75.0 Whiskered Auklet 140 112.7 36.3 60-237 54 85.2 Parakeet Auklet 170 175.4 54.6 96-242 11 36.4

a Conservative estnnate of minimum height of copulating pair based on two body thicknesses, using body thickness measurements of Least and Crested Auklets, and estimates of body thicknesses of Whiskered and Parakeet Auklets based on the relative sizes of the four speaes.

St. Paul Island engaging in this behavior (Fig. In all, we observed 394 copulation attempts 2e). Six pairs of Parakeet Auklets were seen to at sea comprising 166 behaviorally successful engage in a sky-pointing display as an addition and 228 unsuccessful copulations (attempted: to the mutual-court display. This took the form Least 176, Crested 157, Whiskered 35, Parakeet of one member of the pair stretching its head 26; successful:Least 71, Crested 69, Whiskered and neck up and pointing its bill nearly verti- 15, Parakeet 11). Copulation was similar in all cally upwards while vocalizing. four auklet species.The male did not mount the Taking all displays together, the four species female as in other avian species,but instead po- differed in the frequency in which they engaged sitioned himself at the rear of the female, raised in courtship (x*~ = 55.0, P < 0.001: Fig. 2f), himself up with flapping wings, brought his clo- with Whiskered Auklets exhibiting the highest aca up under the female’s tail and in an upright courtship frequency (95.3%, 41143, of observed vertical stance, maintained by vigorous wing pairs engaged in courtship), Crested and Para- flapping and using his feet to steady himself, keet Auklets an intermediate courtship frequen- brought his cloaca up to meet the female’s clo- cy (Crested Auklets: 78.9%, 194/246, of ob- aca. The lift generated by rapid beating of the served pairs engaged in courtship; Parakeet wings appeared to give the male lift to push the Auklets: 67.0%, 1501224, of observed pairs en- female’s cloaca up from the water (Fig. le). By gaged in courtship), and Least Auklets the low- copulating without mounting, males avoided est courtship frequency (55.4%, 2241404, of ob- submerging the female’s cloaca under water dur- served pairs engaged in courtship). For Least ing insemination. Hence, insemination occurred Auklets there was no difference in the frequen- well clear of the sea surface avoiding water cies of any of the courtship displays performed damage to sperm. by pairs at Buldir and those at St. Paul Island Most pairs of auklets engaged in a single cop- (hunch: x2, = 3.4, P = 0.07; pursuit: x2, = 0.6, ulation attempt during any observation period, P = 0.44; neck-twist: x2, = 0.1, P = 0.80; court: but up to five copulation attempts in rapid suc- x*1 = 0.5, P = 0.49). Similarly, there was no cession were observed in focal pairs of Least difference in the frequency of hunch or pursuit and Crested Auklets, and up to three consecutive displays performed by Parakeet Auklet pairs at attempts were seen in Whiskered and Parakeet Buldir and those at St. Paul Island (hunch: x2, Auklets. One Crested Auklet pair engaged in = 0.4, P = 0.54; pursuit: x2, = 0.2, P = 0.70). five behaviorally successful copulations and a However, Parakeet Auklets at Buldir performed Least Auklet pair engaged in three behaviorally more mutual-court displays than those at St. successful copulations in succession. However, Paul Island (x2, = 6.7, P = 0.01). overall there was no difference among the four COPULATION BEHAVIOR speciesin the incidence of single as opposedto All copulations occurred at sea. It is unlikely multiple copulation attempts or single as op- that pairs copulated in their nesting crevices as posed to multiple behaviorally successfulcopu- the majority of Least, Crested, and Whiskered lations (attempts: x*3 = 4.5, P = 0.21; behav- Auklet crevices were too small to accommodate iorally successfulx*3 = 3.3, P = 0.34). a pair in the copulation position used by most There was no difference among the four spe- birds in which the male is positioned over the cies in the likelihood of a pair being behavior- females back (Table 1). ally successful when engaging in one or more 524 F. M. HUNTER AND I. L. JONES copulation attempts during an observation peri- Extrapair copulation attempts were observed od (xz3 = 1.3, P = 0.73). Failure of copulation in all four species of auklet, and unforced be- attempts prior to apparent cloacal contact oc- haviorally successful EPCs occurred in Least, curred for a number of reasons, including male Crested, and Whiskered Auklets (members of failure to achieve or maintain correct position- focal pairs engaging in EPC attempts: Least 33, ing, female avoidance of copulation, and disrup- Crested 12, Whiskered 2, Parakeet 2; members tion by an extrapair individual. Female cooper- of focal pairs engaging in behaviorally success- ation was necessary for a copulation attempt to ful EPCs: Least 3, Crested 2, Whiskered 1). result in a behaviorally successful copulation. EPCs took the form of an extrapair male rapidly Females had to actively maintain their position swimming up to, or flying in and landing close in the water and hold their tails in an upright to a pair, and attempting to copulate with the position for the pair to achieve cloaca1 contact. female. The pair male was usually close by and Females of all four species were observed both in most cases he attempted to disrupt the copu- avoiding a male’s courtship advances and ter- lating individuals and either chase the extrapair minating copulation attempts. They did this by male away or drive his female away from the flying away, swimming forward, diving, or sim- extrapair male. In every case in which the be- ply keeping their tails either flattened against the havior of the focal individuals was recorded af- water surface or submergedunder water. Female ter the EPC attempt, the female remained with Parakeet Auklets terminated a relatively high her original partner (Least 13, Crested 11, Whis- proportion of copulation attempts (36.4%, 4/l 1). kered 1, Parakeet 1). Pair males often attempted Female Crested Auklets caused failure in 27.9% copulations with their partners immediately fol- (17161) of copulation attempts, and female lowing extra-pair copulation attempts (Least: 6 Whiskered Auklets were responsible for failure of 33 [18%] attempted EPCs followed by a pair in 18.8% (3/16). Female Least Auklets caused copulation attempt; Crested: 5 of 12 [42%]; relatively few failures (8.6%, 6/70). Whiskered: 1 of 2 [50%]). Intruding conspecifics disturbed pairs at all stages during courtship and copulation. Disrup- DISCUSSION tions varied from individuals swimming towards Auklets are unusual among seabirds in having a courting pair, to individuals flying in and exclusively aquatic copulation and related court- knocking the male from his copulating position, immediately prior to or during apparent cloaca1 ship behavior at sea. Most of our focal pairs of contact. Least Auklets experienced the highest Least, Crested, Whiskered, and Parakeet Auklets level of courtship disruption, with 54.4% (43/79) engaged in courtship displays during the morn- of courtship bouts being disrupted by one or ing activity period at their staging areas near the more conspecifics. Crested Auklet courtship breeding colony. Pairs also bathed and preened bouts were disrupted in 32.3% (52/161) of cases, while at the staging area at this time, however whereas Whiskered and Parakeet Auklets expe- single individuals appearedjust as likely to en- rienced lower levels of courtship disruption gage in self-maintenance as paired birds. Thus (Whiskered: 10.0%; 3/30, Parakeet: 14.1%, 9/ it appearsthat the main reason for auklets being 64). Disruptions occurring during copulation present in pairs near the staging flock was to were less common for all species (Least: 3.4%, engage in courtship and copulation behavior. 2159; Crested: l.l%, l/87; Whiskered: 4.5%, 11 Some behaviors displayed during courtship at 22; Parakeet: O%, O/14). Disruptions were fol- sea were similar to land-based courtship dis- lowed by one or both members of the pair driv- plays. The hunch display occurs in both Least ing the intruder away, or one member of the pair and Crested Auklets on land, and the neck-twist driving the other member of the pair away from and ruff-sniff displays occur in Crested Auklets the intruder. on land (Jones and Montgomerie 1992, Jones One copulation attempt involving a pair of and Hunter 1993). Little is known of Whiskered Least Auklets was terminated by a Glaucous- and Parakeet Auklet courtship behavior on land winged Gull (Lanes gluucescens).The female of because of the Whiskered Auklet’s nocturnal the pair dived secondsbefore the attack and the disposition and the Parakeet Auklet’s dispersed male was caught by the predator and swallowed nesting and crepuscular habits. Consequently, it whole. is not known whether the Whiskered Auklet’s AUKLET COURTSHIP AND COPULATION 525 head-bob or the Parakeet Auklet’s mutual-court males are able to perform forced copulations display occur on land. (Mineau and Cooke 1979, Bums et al. 1980). With the exception of pursuit behavior, all dis- Forced copulations were not observed in any of plays focused on the head region of the display- the four species of auklets. Instead, female co- ing birds, where the ornaments of all four spe- operation was necessary for a behaviorally suc- cies are situated. Therefore, it seems probable cessful copulation to be achieved. that the displays function to highlight or draw In this study we defined a pair as two birds attention to ornaments and that individuals en- maintaining close proximity over a period of at gaging in courtship displays assesstheir part- least 2 min. We justified this definition on the ner’s ornaments. Crested Auklets’ mff-sniff and basis of observations of known pairs leaving the neck-twist displays were conspicuous in focus- talus together and engaging in behavior identical ing specifically on the back and sides of the to that of our focal pairs. However, in the ab- neck. Crested Auklets have a distinctive citrus- sence of individual identification of focal birds, like plumage odor which is very noticeable on we could not be certain that all pairs defined in the nape and neck feathers (pers. observ.) Al- this way were breeding partners or that extrapair though it was not possible to determine whether males were not in fact breeding partners of the the birds were actively smelling their partners’ focal female. Following an EPC attempt, fe- neck feathers during the ruff-sniff or neck-twist males always returned to the original focal male displays, it is possible that this odor is a sexually and these males often engaged in copulation at- selected trait. tempts so that females copulated with two dif- All auklet displays were carried out by both ferent males within minutes of each other. This sexes. The neck-twist, head-bob, and mutual- supportsthe idea that extrapair males were not court displays involved simultaneous displaying the female’s breeding partners and that sperm by both sexes,whereas the pursuit and mff-sniff competition is a feature of auklet mating sys- displays were performed at different times by tems. Nevertheless, studies of individually male and female. Sexually monomorphic court- marked birds have shown the occurrence of un- ship behavior is consistent with similar oma- forced EPCs between two birds in the temporary ments being expressed in both sexes and with absence of the partner of one of them (Hatch mutual mate choice. Mutual mate choice has 1987, Hatchwell 1988, Hunter et al. 1995). Such been experimentally demonstrated in Least and unforced EPCs tended to be behaviorally indis- Crested Auklets (Jones and Montgomerie 1992, tinguishable from pair copulations (E M. Hunter, Jones and Hunter 1993), and, on the basis of pers. observ.). If a female and extrapair male mutual courtship behavior and mutual expres- consorted in this manner in the present study, sion of ornaments, it is likely to occur in Whis- they would not have been identified as an extra- kered and Parakeet Auklets. pair partnership. However, rather than artificially Three courtship behaviors, hunch, pursuit and inflating the importance of EPCs, this would mutual court, were observed in all four species tend to result in underestimation of the incidence of auklet, whereas other displays such as the of extrapair activity. ruff-sniff display of the Crested Auklet or the All copulations that we observed in this study head-bob of the Whiskered Auklet were unique occurred at sea. During over 1,500 hours of ob- to one species.This partitioning of displays may servations of Least and Crested Auklets on land have been useful, at some time in the past, in at the Main Talus breeding colony on Buldir facilitating speciesrecognition. Although auklet during both morning and evening activity peri- species are morphologically quite distinct at the ods (Jones and Hunter 1993, 1998, in press), no present time, it is possible that these displays copulation was ever observed either on the sur- evolved during speciation events by helping in- face of the talus or in protected areas below or dividuals to avoid copulating with speciesother behind boulders. Furthermore, the majority of than their own. This is consistent with the ob- Least, Crested, and Whiskered Auklet crevices servation that the two most closely related auklet were too small to accommodate a mounted pair, species (Crested and Whiskered) have the most so it is unlikely that many pairs copulated in distinct courtship display and elaborate morpho- their nesting crevices. Hence, it appears that logical adornments. auklet copulations must occur mainly or exclu- In species which have an intromittent organ, sively at sea. 526 F. M. HUNTER AND I. L. JONES

The weight of a male bird mounting a female ulate on land than at sea. So the crucial question in water is likely to cause the female’s cloaca to is why do auklets and puffins, which lack intro- become submerged. If sperm were inseminated m&tent organs, copulate at sea instead of on land by the usual avian method of cloaca1contact, the like other alcids? We proposeseveral hypotheses male’s sperm would likely be washed away or to explain auklets’ propensity for aquatic mat- damaged by the surrounding water. Male water- ing. fowl have an intromittent organ so sperm can be (1) Predation avoidance. At many colonies, safely inseminated inside the female even if her auklets suffer frequent attacksby gulls and other cloaca becomes submerged when the male is avian predators suggesting that birds engaging mounted (Lake 1981, Briskie and Montgomerie in copulations on land might be vulnerable to 1997). Auklets, which copulate at sea but do not predation. Pairs mating at sea had an excellent have an intromittent organ, have an alternative view of approaching avian predators and es- way of overcoming the potential problem of clo- caped attacks quickly by diving. One pair of acal submergence and water damage to sperm. Least Auklets in our study suffered predation Male auklets did not mount females when cop- while copulating at sea but predation pressure ulating at sea. By remaining unmounted and in- on land apparently was greater than that occur- stead using rapid flapping of the wings to bring ring at sea. If copulation on land carries a pre- himself into the correct position for cloacal con- dation risk, it would be predicted that auklets tact, male auklets avoided submerging their part- would copulate at protected sites within the talus ners and running the risk of causing water dam- where they would be immune from gull preda- age to sperm during insemination. This unortho- tion. We have made many observations of Least dox system of copulating may be the only way and Crested Auklet courtship behavior on the males can achieve sperm transfer by cloaca1 talus (Jones and Hunter 1998, in press) during contact without submerging the female. The which protected sites beneath overhanging or unique alcid shape with the legs set far back on jumbled boulders were visible, but we have nev- the body and the wings compact and powerful, er seen a pair of auklets copulating on land. Nor might facilitate this form of copulation, although does it appear possible for many pairs to copu- similarly shaped alcids such as the murres and late in the protection of their nesting crevices as penguins copulate by mounting on land. most of these were too small to accommodate a Within the Alcidae, the practice of copulating pair of copulating auklets. Furthermore, even if at sea appears to be phylogenetically con- predation pressure on the talus surface was a strained. There are two major groups within the problem for Least and Crested Auklets, it is not Alcidae, one containing the puffins and Rhinoc- likely to be a problem for the Whiskered Auklet eros Auklet (Cerorhinca monocerutu, Fratercu- which is active on land only at night, yet this lini) and the true auklets (Aethiini), the other species still copulates at sea. Thus, predation containing the murres, Razorbill (Alca tordu) pressurealone does not appear to provide an un- and Dovekie (Alle de, Alcini), the guillemots equivocal explanation for aquatic copulation. (Cepphini), the synthliboramphine murrelets (2) Harassment and EPC avoidance. Females (Synthliboramphini), and the brachyramphine copulating on land within the colony site might murrelets (Brachyramphini; Friesen et al. 1996). find it more difficult to avoid harassment and Copulation location has been established for 16 forced EPC attempts by extrapair males. Ha- of the 23 extant alcid species.Seven of the eight rassment which carries costs of wasted time and species known to copulate primarily or exclu- potential injury to the female or her unlaid egg sively at sea all belong to the first group, where- may come in the form of forced extrapair cop- as the eight remaining species known to copu- ulation attempts which occur in Pigeon Guille- late primarily or exclusively on land all belong mots (Cepphus columbu), Common Murres to the second group. The one exception to this (Uriu uulge), and Razorbills, all of which cop- pattern is the Marbled Mm-relet which phylo- ulate on land (Drent 1965, Birkhead et al. 1985, genetically belongs to the murre-guillemot-mur- Wagner 1991). Behaviorally successful forced relet group but has been recorded copulating EPCs have been reported in Common Murres most often at sea (Gaston and Jones 1998). (B&head et al. 1985, Hatchwell 1988). Under It would appear that for a species that lacks normal circumstances it seems unlikely that an intromittent organ it would be easier to cop- males in species that lack intromittent organs AUKLET COURTSHIP AND COPULATION 527 can force copulations on females (Fitch and late at their nest sites because of the need for at Shugart 1984). Indeed, Wagner (1991) argues least one pair member to be present at all times convincingly that female Razorbills are in con- to defend the nest site from conspecifics. Nev- trol of copulation success.However, male Com- ertheless, the question of whether cleanliness mon Murres have adopted a strategy to over- plays a role in determining copulation location come uncooperative females and achieve suc- remains open. A detailed comparative study of cessful forced EPCs. Males sometimes engage cleaning behavior and pathogen transmission in in multiple-male EPC attempts, and when four relation to copulation location in alcids would or more males are involved, the female can be be required to resolve this question. pinned down enabling some males to achieve (4) Female testing of males. If it is hard for a cloacal contact (Birkhead et al. 1985). By being male to achieve a successful copulation at sea, receptive only at sea, females retain the option he may be honestly signaling his quality to the of escape by diving at any time and thus can female by copulating in this manner. By accept- avoid unwanted copulations. Similar avoidance ing copulation attempts only at sea, the female of unwanted copulations is seen in Atlantic Puf- may gain by having the opportunity to test the fins (Creelman and Storey 1991). It would be quality of the male. predicted that if multiple male groups of auklets (5) Nonadaptive. Despite the apparent com- could force EPCs on females on land, they plications of aquatic copulation, it may not be would do so during those activity periods in any more difficult for an auklet pair to copulate which males and females court at the breeding at sea than on land. It may be that the ancestor colony on the talus surface, but no copulation of the auklets and puffins copulated at sea and was ever observed on land. Female auklets on a thus that modern species continue to do so in flat boulder may always be able to move away the absence of any opposing selective force. from males attempting EPCs, unlike female In summary, unlike most other bird species, murres pinned against a cliff face. At sea, female and even unlike most seabird species, auklets auklets would be able to escape harassment by court intensely and copulate mainly or exclu- diving and reduce their susceptibility to distur- sively at sea. On the basis of our data, we were bance by avoiding areas of high bird density. unable to unequivocally discriminate among Mating pairs often frequented the periphery of possible explanations which could play a role in the staging flock where relatively fewer conspe- favoring aquatic copulation. The question of cifics were present. Hence, it appears unlikely why these birds copulate at sea remains an enig- that aquatic copulation in auklets is based on a ma that deserves further attention. female strategy to avoid forced copulations, but pairs may mate at sea to avoid harassment by ACKNOWLEDGMENTS conspecifics. Many thanksto VernonByrd, Daniel Boone, and Jeff (3) Cleanliness. Individuals spending time in Williams of the U.S. Fish and Wildlife Service for uro- their crevices and under the talus are prone to viding logistical support and permission to carry&out auklet researchin the Alaska Maritime National Wild- getting their under-tail coverts, and therefore the life Refuge. Many thanks to Art Sowls for setting up area around the cloaca, soiled with feces and logistical details, and Steve Zimmerman and the Na- mud. During copulation on land, dirt and asso- tional Marine Fisheries Service for providing accom- ciated micro-organisms would likely be trans- modation during the St. Paul Island work. We are ferred onto the cloaca of a copulating partner grateful to Arctic King Fisheries and the captain and crew of the F/T Resolutefor transportingus to Buldir (Sheldon 1993). Aquatic copulation with a part- Island in 1995 early enough to allow this work to be ner who has washed on the water surface before carried out, and to Kevin Bell and the crew of the M/ copulating would likely reduce the chances of V Tiglh for safe transportand excellent logistical sup- pathogen transmission. Auklet pairs bathed and port in all years of the study. We thank Jim Briskie, Tony Gaston, and Anne Storey for helpful comments preened intensely while in the staging flock on the manuscript. The research was funded by Na- which supportsthis idea. However, the question tional Geographic, the NERC, U.K. (FMH), and the arises as to why birds did not wash and then NSERC, Canada (ILJ). return to land to copulate. Dirt is likely to pose a similar problem for murres and Razorbills LITERATURE CITED which nest and copulate on feces-covered cliff BIRKHEAD, T R., S. D. JOHNSON,AND D. N. NETTLE- ledges. Murres at least are constrained to copu- SHIP.1985. Extra-pair matings and mate-guarding 528 F. M. HUNTER AND I. L. JONES

in the Common Murre (Uris aalge). Anim. Behav. JONES, I. L. 1993a. Crested Auklet. In A. Poole, l? 33:608-619. Stettenheim, and E Gill [eds.], The birds of North BRISKIE,J. V., ANDR. MONTGOMERIE.1997. Sexual se- America, No. 70. The Academy of Natural Sci- lection and the intromittent organ of birds. J. Avi- ences, Philadelnhia, and the American Omitholo- an Biol. 28:73-86. gists’ Union, Washington, DC. BURNS,J. T, K-M. CHENG, AND E MCKINNEY. 1980. JONES.I. L. 1993b. Least Auklet. In A. Poole. I? Stet- Forced copulation in captive Mallards. I. Fertil- tenheim, and E Gill [eds.], The birds of North ization of eggs. Auk 97:87%X79. America, No. 69. The Academy of Natural Sci- BYRD G. V., R. H. DAY, AND E. I? KNUDTSON1983. ences, Philadelphia, and the American Omitholo- Patterns of colony attendance and censusing of gists’ Union, Washington,DC. auklets at Buldir Island, Alaska. Condor 85:274- JONES,I. L., AND E M. HUNTER.1993. Mutual sexual 280. selection in a monogamousseabird. Nature 362: CREELMANE., ANDA. E. STOREY. 1991. Sex differenc- 238-239. es in reproductive behavior of Atlantic Puffins. JONES,I. L., AND E M. HUNTER.1998. Heterospecific Condor 93:390-398. preferences for a feather ornament in Least Auk- DANIELS,D., J. HEATH,AND S. V. STEVENAGE. 1994. lets. Behav. Ecol. 9: 187-192. Courtship behavior in offshore Kittiwake Rissa JONES,I. L., ANDI? M. HUNTER.In press.Experimental triductyla flocks prior to the breeding season.Sea- evidence for mutual inter- and intra-sexual selec- bird 16:15-21. tion favouring a Crested Auklet ornament. Anim. DARWIN,C. 1871. The descentof man and selectionin Behav. relation to sex. Murray, London. JONES,I. L., ANDR. MONTGOMERIE.1992. Least Auklet ornaments:do they function as quality indicators? DRENT,R. H. 1965. Breeding biology of the Pigeon Behav. Ecol. Sociobiol. 30:43-52. Guillemot Cepphuscolumba. Ardea 53:99-160. LAKE, F?E. 1981. Male genital organs, p. 1-61. In A. FRIESEN,V. L., A. J. BAKER,AND J. E PIA=. 1996. S. King and J. McLelland teds.], Form and func- Phylogenetic relationships within the Alcidae tion in birds. Academic Press, London. (: Aves) inferred from total mo- MCIQNNEY, E, S. R. DERRICKSON,AND F.’ MINEAU. lecular evidence. Mol. Biol. Evol. 13:359-367. 1983. Forced cooulation in waterfowl. Behaviour GASTON,A. J., ANDI. L. JONES.1998. The . Ox- 86:250-294. I ford Univ. Press, Oxford. MINEAU, I?, AND I? COOKE.1979. Rape in the Lesser HATCH,S. A. 1987. Copulation and mate-guardingin Snow Goose. Behaviour 70:280-291. the Northern Fulmar. Auk 104:450-461. SHELDON,B. C. 1993. Sexually transmitted diseasein HATCHWELL,B. J. 1988. Intraspecificvariation in extra- birds: occurrence and evolutionary significance. pair copulation and mate defense in Common Phil. Trans. R. Sot. Lond. B. 339:491-497. Guillemots Uris aalae. Behaviour 107:157-185. WAGNER,R. H. 1991. Evidence that female Razorbills HUNTER,E M., T. BURKE,“ANDS. E. WATTS.1992. Fre- control extra-pair copulations. Behaviour 118: quent copulation as a method of paternity assur- 157-169. ance in the Northern Fulmar. Anim. Behav. 44: ZUBAKIN,V. A. 1990. Some aspectsof the nesting bi- 149-156. ology and social behavior of the Crested Auklet HUNTER,E M., G. D. MILLER,AND L. S. DAVIS. 1995. (Aethia cristatellu), p. 9-13. I?t A. Kondratiev Mate switching and copulation behaviour in the [ed.], Study of colonial seabirds of the USSR. AdClie Penguin. Behaviour 132:691-707. Acad. Sci. USSR, Magadan, Russia.