64 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

POPULATION BIOLOGY AND ADULT BEHAVIOR OF AROTA ()

JAMES A. SCOTT 60 Estes Street, Lakewood, Colorado 80226

The purpose of this paper is to describe adult behavior (mate­ locating, mating, feeding, oviposition, and basking), dispersal, and pOpu­ lation parameters (especially lifespan) of L. arota Boisduval. A mark-recapture study was conducted in 1969 at the mouth of Rouch Gulch (Spring Creek), Fremont County, Colorado. L. arota was also studied in 1971 one km east of Smith Creek Campground, Custer County, Colorado, and at Little Fountain Creek, EI Paso County, Colorado. These sites were mountain gulches or streams with abundant larval host, spp. METHODS Number of matings per female was determined by counting spermato­ phores (Burns, 1968). For study of movements and estimation of popu­ lation parameters, were marked individually using the method of Ehrlich & Davidson (1960). They were captured with a net, marked, and immediately released at the site of capture. Analysis of population movements. The following method allows direct comparison between the sexes and between species, determination of change of movements with age, and separation of the velocity and distance aspects of movements. On a map of the movements of each recaptured individual, distances in mm between each two captures are measured, and called d1 between first and second captures, d 2 between second and third, etc. Total distance moved by the individual, D, is the sum of the d's. Range, R, an estimate of dispersal, is distance between the two most distant capture points. Time in days between first and second capture is called t1, etc. T is total time between first and last capture. Velocities are defined Vi = d;/ti and V = D/T. Midpoint age between captures is determined by finding the age midway between two captures after calling the first capture day O. Correlations between distance or velocity and midpoint age determine whether movements change with age. Midpoint age is used rather than age at start or end of a period between captures because the time between captures differs. Jolly's stochastic method was used to estimate population size, survival rates, and number of new individuals joining the population (Jolly, 1966). The method of Cook et al. (1967) was used to obtain expected lifespan from average survival rate. VOLUME 28, NUMBER 1 65

RESULTS Mate-locating behavior. To locate females, males perch on branches of shrubs and trees 1 to 2 m above the ground. Males start perching at about 0715 and actively perch until noo. Males gradually stop perching between noo and 1200, and rarely perch after 1200. In the afternoon, males mainly visit flowers, or are quiescent on shrubs. Perching males sit on a branch or leaf and dart out at passing objects. Most such objects are other males, whereupon the two males fly about each other for a few seconds, then each returns to or near the same perch as before. One male was observed to perch on a one-meter length of oak branch for two hours dming which time he made almost 100 short flights at passing objects. Other males sometimes flew short distances between investigative flights. Males chase a narrow size range of passing objects, so that almost all of the chases are toward other L. arota. Males perch in small clearings in many different topographic situations. A clearing likely to have perching males is about 3-5 m. in diameter, rea­ sonably level, and surrounded by tall trees or steep hills or both. Males were observed to perch in clealings that were in a small valley bottom, on a hillside along an irrigation ditch, and on a ridgetop. More males perch in valleys than hillsides or hilltops because more suitable clear­ ings are found there and the larval hostplant is usually more abundant there. Mating. If a female flies past a perched male, the male darts after her and flies about 14 cm below her for several meters, then the female lands on a branch and the male lands behind. The male then usually flicks his wings by holding them about 60° from the vertical (about 120° from each other) and vibrating them at small amplitude (only 1-2 mm amplitude for each wing). Sometimes the male vibrates the wings only once or twice per second, and other times he vibrates his wings rapidly (about 10 times per second), occasionally with the vibrations clustered into groups. These two types of wing flicking occurred about equally often, but only one type was seen in anyone courtship, except for one courtship in which the male vibrated his wings once per second, then drew closer to the female and vibrated his wings about 10 times per second. Other males did not flick their wings at all, but merely walked to the female and attempted copulation. \Vhether or not the male flicks his wings, if the female remains quiescent the male crawls along­ side and bends his abdomen either right or left to attempt copulation. In the two completed copulations observed, the female was quiescent during courtship and mating. Unreceptive females flap the wings almost full stroke about 10 or more times per second for about 2-5 seconds, while 66 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

TABLE 1. Dispersal data for Lycaena arata. N = sample size. Dispersal param­ eters are defined in methods section.

Dispersal Parameter Males N Females N Number Marked 107 94 Number Recaptured 53 37 Average T (days) 2.56 53 1.91 37 Average t. (days) .94 144 .84 85 Average R (meters) 15 53 29 37 Average D (meters) 16 53 29 37 Average d, (meters) 6 144 13 85 Average V (meters per day) 11 53 18 37

Average Vi (meters per day) 5 144 9 85 sitting. In most courtships the female was previously mated, and she performed this "rejection dance" when the male crawled up to her; the male flew away then or after subsequent rejection dances. Male wing flicking may cause the female to become quiescent and receptive, because if the female is already quiescent, the male usually does not flick his wings. Courtship and mating occur at the same time as male perching. Copulating pairs were found at 0805, 0920, 1121 and 1200, all in valley bottoms by Scott, and at 0930 and 1326 (Oakley Shields, pers. comm.). 1326 is after the normal perching period, and that observation may represent the mating of a late perching male, or perhaps the mating was initiated during the normal perching period and the pair remained joined until observed (copulation of butterflies lasts rarely up to 30 hours). Nineteen courtships were observed from 0815 to 1036, and one was observed at 1121. Females rarely mate more than once. Of 60 females dissected (caught several weeks after the species had first appeared), 15 were virgin, 44 had mated once, and only 1 had mated twice. Many virgin females were found in the afternoon, indicating that many females wait until the day after emergence to mate. Movements. The mark-recapture study was carried out from 30 July to 8 August 1969. Rouch Gulch is a small, dry (except after rains) gulch opening into the Arkansas River, within the pinyon-juniper belt. The larval host, Ribes leptanthum, and adult nectar sources were scattered along the bottom, where the recapture study was carried out from the Arkansas River to 300 m. up the gulch. Nearby gulches were sampled to detect dispersal. The proportion of recapture (Table 1) was higher for males, probably because males disperse less than females. Although TABLE 2. Population parameters estimated from multiple recapture data using the stochastic model of Jolly (1966). Alpha-pro- -<0 portion of marked animals; M-total marked population; N-total population; Phi-probability of survival; B-number of new t"' joining the population; SE-standard error. C ~ ~ Day Alpha M N + 1.96 SE Phi + 1.96 SE B + 1.96 SE to vCXJ Males Z July 30 .959 + .394 c ~ 31 .2581 43.13 167.1 + 112.9 .662 + .275 13.1 + 69.4 IJl ~ Augllst 1 .3538 1,.3.77 123.7 + 45.4 .997 + .758 91.6 + 90.2 i:C ...... 2 .3836 80.50 ~09.9 + 164.6 .164 + .159 68.7 + 46.4 5 .1406 13.50 96.0 + 53.5 .593 + .548 47.7 + 63.0 6 .1111 8.00 72.0 + 70.4 7-8 .1818

Females July 31 .374 + .242 August 1 .0833 14.23 170.7 + 160.8 1.076 + .767 79.8 + 187.8 2 .2283 59.40 260.2 + 200.6 5,6,7 .1000

Both Sexes Combined July 30 1.157 + .476 31 .1304 57.87 443.7 + 311.1 .489 + .177 24.3 + 146.6 August 1 .2389 57.62 241.2 + 82.8 1.022 + .528 228.4 + 160.7 2 .2970 138.56 466.6 + 253.5 .264 + .224 209.0 + 178.3 5 .1311 40.00 305.0 + 228.4 .438 + .514 105.4 + 155.4 6 .0909 17.50 192.5 + 205.6

7-8 .1087 0:> -1 68 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

TABLE 3. Flowers and other fluid sources visited by L. arota at the three study sites.

Rouch Little Gulch Smith Fountain Creek Species Color Creek Male Female Cmpgd. Solidago occidentalis yellow few 236 158 Pericome caudata yellow 52 75 Eriogonum iamesi whitish -yellow few 30 64 Heterotheca villosa yellow 5 9 Chrysothamnus naltseOSllS yellow 6 5 H elianthus pumilus yellow 2 2 Allium sp. white 1 Rubus parviflorus'- 5 3 mud 1 Rudbeckia laciniata yellow many M elilotus alba white many Aster novae-angliae bluish white many Apocynum sp. white many Clematis sp. cream white few common Achillea sp. white few

1 Juices of the blue-black berries. the times between captures for females were less than those for males, the distances moved were much greater for females, resulting in female velocities almost twice those of males. Most individuals of both sexes did not move at all between recaptures, so that the averages of movement statistics in Table 1 are low. Some individuals move considerable dis­ tances, however. Maximum distances moved (ranges) for males were 83, 84, 92, 93, 168, and 214 meters, and for females 82, 83, 85, 92, 94, 169, and 186 meters. Population parameters. The population size in the study area at Rouch Gulch, which is about 300 m. by 100 m., was about 400, declining to 200 at the study's end (Table 2). Females were slightly more common than males since males emerge several days before females and the mark­ recapture study was conducted after the peak emergence of males. The number of new animals joining the population was fairly high, about 60 per day for males and 80 for females. The number of animals joining the population and the survival rates are due mostly to emergence and deaths respectively, because few individuals dispersed out or into the area. The 0.164 survival rate for males is low because it represents the survival over a three day period. The average survival rate for males and corresponding expected lifespan was 0.763 (3.7 days) using VOLUME 28, NUMBER 1 69 looj NUMBER 90 80 OF 70 60 VISITS :~1 TO 30 20 FLOWERS 10 O~--~~~~II-rIlTl~~ ~cQ&U'/~{r.~~--&~ 00°0°00000000000 000000

TIME OF DAY Fig. 1. Number of visits to flowers at various times of day for both sexes of L . awta. method 1, and 0.752 (3.5 days) using method 2 (methods of Scott, 1973). The rate for females was 0.790 (4.2 days) using method 1 and 0.725 (3.1 days) using method 2. The potential lifespan is much longer: males have lived at least 8, and females at least 6, days. The survival rates for both sexes were undoubtedly decreased by extensive predation by robberflies and ambush bugs (see below) and by the weather, which was very hot especially toward the end of the study when the survival rate declined slightly. Feeding. Both sexes very often feed at flowers (Fig. 1). Flower visits were recorded for 336 males and 296 females. Before 1l00, when males are perching, males visit flowers less than females, but after 1100 males visit flowers more than females: 32 males and 58 females were recorded before 1l00; 304 males and 238 females after 1l00. Both sexes prefer yellow and white flowers but feed on different species at different locali­ ties (Table 3). Males preferred Solidago occidentalis at Rouch Gulch slightly more than females, and females preferred Peri come caudata 70 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY and Eriogonum iamesi more than males, but this difference seemed mainly because Solidago was concentrated at the mouth of the gulch (where males were more abundant) whereas the other plants were more widely distributed. The preference for yellow flowers provided ambush bugs with many meals (see predation below). Oviposition. Females oviposit on twigs, bark, or dead leaves, on or under bushes of Ribes species, especially the subgenus Grossularia. Many ovipositions on R. leptanthum were observed at Rouch Gulch and Little Fountain Creek. Two ovipositions on R. montigenum were ob­ served at Williams Canyon, EI Paso County, Colorado. John Emmel (pers. comm.) found ova on R. leptanthum near Glenwood Springs, Garfield Co., Colorado. Emmel et al. (1970) found larvae on R. velutinum in Nevada which were raised to adults. Gunder (1930) found larvae on Ribes gracillimum in southern California. Females often spend an hour or more on the same or a nearby Ribes bush alternating oviposi­ tion about every five minutes with basking, "hindwing rubbing," and sitting. Females oviposit in the center of a bush just as often as on the outer branches. Nine eggs were laid on rough bark of the thicker branches, 12 were laid on sides of smooth twigs of the thinner branches, and 2 were laid on dead leaves of two different dicotyledons underneath Ribes bushes. Females oviposit during the warmer hours of the day from 0900 to at least 1430. The eggs undergo diapause and do not hatch until the next year. Thermoregulation. Both sexes bask by holding the wings about 60 0 from vertical, 120 0 from each other, the same position as in male courtship flicking. Individuals orient their body in any position which brings the wings somewhat perpendicular to the sun, but usually face away from the sun. Both sexes bask more often in morning and late afternoon. A few individuals basked while feeding, but basking is more frequent between oviposition and "hindwing rubbing." Roosting. One male was found on the leaf of an oak tree (Quercus gambellii) at 1708, at Smith Creek Campground, Custer County, Colorado. Predation. Considerable predation was observed. Robberflies (Asil­ idae) captured 4 males and 2 females, and missed others. Ambush bugs (Phymatidae) caught one male on Solidago flowers and 1 male and 3 females on Pericome flowers. A crab spider (Thomisidae) on a composite species caught a female at Little Fountain Creek. L. arota (and all other Lycaenidae except Riodininae) move the hind­ wings forward and back, the left and right wing moving in opposite directions, a behavior which I call "hindwing rubbing." Many males VOLUME 28, NUMBER 1 71 and females did this at all times of day. Basking and hindwing rubbing often follow each other, but do not occur often at flowers. L. arota has 2 mm tails on the hindwing, but does not have a conspicuous eyespot next to the tail, so hindwing rubbing mayor may not serve to divert bird attack away from the body by drawing attention to the antenna-like tail, as hypothesized for other Lycaenidae (see Wickler, 1968). One male and two females had both hindwings truncated in the manner which may indicate bird or lizard attack.

DISCUSSION L. arota differs from L. xanthoides Boisduval and many other Lycaena in that mating occurs only in early morning. The preference of L. arota for perching in small clearings contrasts with the open perching sites chosen by L. xanthoides; these perching sites correspond to the usual habitats of the two species, mountain foothills for L. arota and flat open areas for L. xanthoides. Population movements of L. arota are small, but several factors combine to counteract the weak dispersal: ( 1) females disperse much more than males because their velocities are greater and they live longer; (2) the dispersal of some individuals is much greater than that of most others; (3) larval hostplants are rather scattered, resulting in more continuous populations than if the plants were aggregated and these aggregations widely separated. L. arota is found almost everywhere in southern Colorado at least between 1500 and 2200 m. in the "Vet Mountains and Arkansas River Canyon; and (4) the species is usually very abundant, so that more individuals will move over long distances than if the species was less abundant. Small peripheral populations could therefore be swamped by a few immigrants from a larger population. The species has four named sub­ species, but the amount of geographic variation is small considering the size of the range and the many geographic barriers therein. Courtship is very similar among several species of Lycaena. Wing flapping of females seems to be a rejection dance in L. xanthoides, L. gorgon Boisduval, and L. helloides Boisduval. L. xanthoides males flap the wings with wide amplitude and hover behind or over the female, rarely flapping while sitting behind her. The courtship of L. gorgon appears identical to that of L. xanthoides, and the courtship of L. helloides is very similar to that of L. xanthoides except that male helloides vibrate the wings with slightly less amplitude while holding them about 30° above horizontal. Other aspects of behavior which are similar in L. arota and L. xanthoides include basking, feeding behavior (L. arota may be more restricted to yellow flowers), and "hindwing rubbing." 72 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

SUMMARY Males perch from 0715 to 1100 on shrubs or trees in small clearings and dart out at passing objects in search of females. Mating usually involves male wing flicking. Unreceptive females flutter their wings until the male departs. Adults are very sedentary, although some individuals move several hundred meters. Both sexes live an average of only 4 days, probably because of hot weather and extensive predation. Both sexes often feed, usually on yellow or white flowers. Females oviposit mainly on Ribes branches. ACKNOWLEDGMENT I thank Jerry A. Powell for improving the manuscript.

LITERATURE CITED BURNS, J. M. 1968. Mating frequency in natural populations of skippers and butterflies as determined by spermatophore counts. Proc. Nat. Acad. Sci., Washington, D.C. 61: 852-859. COOK, L. M., L. P. BROWER & H. J. CROZE. 1967. The accuracy of a population estimation from multiple recapture data. J. Anim. Ecol. 36: 57-60. EHRLICH, P. R. & S. E. DAVIDSON. 1960. Techniques for capture-recapture studies of populations. J. Lepid. Soc. 14: 227-230. EMMEL, J. F., O. SHIELDS & D. E. BREEDLOVE. 1970. Larval foodplant records for North American rhopalocera. Part 2. J. Res . Lepid. 9: 233-242. GUNDER, J. D. 1930. Butterflies of Los Angeles County, California. Bull. So. Calif. Acad. Sci. 29: 39--95. JOLLY, G. M. 1966. Explicit estimates from capture-recapture data with both death and immigration-stochastic model. Biometrika 52: 225-247. SCOTT, J. A. 1973. Convergence of population biology and adult behaviour in two sympatric butterflies, Neominois ridingsii (Papilionoidea, Nymphalidae) and Amblyscirtes simius (Hesperioidea, Hesperiidae). J. Anim. Ecol. (in press). WrcKLER, W. 1968. Mimicry in Plants and Animals. McGraw-Hill, New York. 160 p.

REPORT OF THE CAPTURE OF AN ADDITIONAL HYBRID BETWEEN LIMENITIS ARTHEMIS ASTYANAX AND L. ARCHIPPUS (NYMPHALIDAE) On 15 October 1972, a wild male interspecific hybrid between Limenitis arthemis astyanax (Fabricius) and L. archippus ( Cramer) [form rubidus Strecker] was captured in Durham County nine miles south of Durham, North Carolina. This specimen was captured approximately one-half mile from the site at which a similar hybrid was found two years previously on 10 October 1970 (Platt & Green­ field 1971, J. Lepid. Soc. 25: 278- 284). The specimen presently is in the collection of A. P. Platt at the University of Maryland Baltimore County. The recent hybrid (Fig. 1) is in good condition, showing little evidence of being flight-worn. It was quite vigorous in flight and eluded capture several times before being netted. This specimen closely resembles the lab-reared dark morph depicted by Platt & Greenfield (1971); it entirely lacks all remnants of the medial partial white banding (an archippus character). On the other hand, the previously caught wild hybrid from Durham was battered, and clearly showed traces of the white