OBSERVATIONS ON BREEDING BEHAVIOR OF PA CHYDIPLAX LONGIPENNIS

Pachydiplax longipennis Burmeister is a medium sized dragonfly found throughout the United States. Clifford Johnson (962) has described the general pattern of breeding behavior of this species, in North Carolina, Virginia and Texas. In this paper I report my observations of three populations of this species in Massachusetts. While territorial and mating behavior is similar to that described by Johnson, I am able to extend his findings in some areas. In addition, I describe the coloration of this species in ultraviolet light and speculate as to the behavioral significance of the high reflectancy of areas of blue pruinescence.


INTRODUCTION
Pachydiplax longipennis Burmeister is a medium sized dragonfly found throughout the United States. Clifford Johnson (962) has described the general pattern of breeding behavior of this species, in North Carolina, Virginia and Texas. In this paper I report my observations of three populations of this species in Massachusetts. While territorial and mating behavior is similar to that described by Johnson, I am able to extend his findings in some areas. In addition, I describe the coloration of this species in ultraviolet light and speculate as to the behavioral significance of the high reflectancy of areas of blue pruinescence.

I/IETHODS
Observations were made at three ponds near Boston, Massachusetts, during the period June-August, I974

MALE-MALE INTERACTIONS
From the time o their arrival (approx. o a.m.) until their departure (approx. 4 p.m.), the males observed deended a deqned territory as described by Johnson (962). The. "threat display" (Johnson, 9.62) was also requently seen. l'ollowing this. display, I observed that the two males engage in urther agonistic behavior. One male pursues the other in a horizontal direction with abdomen raised while the leading insect flies with abdomen lowered (Fig. a).
Psyche [March degrees. Immediately, one male returns directly to defend the territory while the defeated male returns many seconds later to rest or does not return to the area at all. Unfortunately, the flight speed was so great, and the distance so far, that I was unable to determine which male claimed the territory, or whether either regularly did, but I would hypothesize that the lower and therefore faster male usually dominates. As Johnson (962)  PERCHINO POSITION The perching position of Pachydiplax longipennis is probably determined by several factors including wind, sun intensity, air temperature, the kind of perch, and the behavior of other dragonflies. Under typical summer daytime conditions when the wind speed is low, a male in an open situation exposed to full sunlight, usually aligns his body along the length of the perch. He raises his abdomen only a few degrees and brings his wings forward (Fig. b). Every o to 60 seconds he leaves his perch to patrol but returns directly if undisturbed by intruders. Very rarely did any male raise his abdomen to the extent illustrated by Johnson (962: Fig. 2). On windier days, males keep their wings ully outstretched and at right angles to the prevailing wind. Only on hot calm da.ys, when the air temperature rose above 3oC, did perching males raise their abdomens higher.
My observations, coupled with Johnson's (r962), suggest that male perching behavior is influenced by at least our actors including thermoregulation, aerodynamics, predator avoidance, and territorial display. It is interesting to note that whereas Johnson (I962) and Williamson (r9oo) reported that males generally perch with their abdomens directed upwards, this behavior was rarely seen in Massachusetts populations. Johnson and Williamson's observations were made on sunny days, at stations where daytime temperatures are typically 2-4C higher than those prevailing in Massachusetts.
This suggests that this behavior may be temperature dependent. As male-male territorial interactions were otherwise ully developed in the Massachusetts populations, I suspect that the abdomen orienting behavior may be primarily thermoregulatory in unction rather than purely or territorial display as concluded by Johnson (1962).

IATING AND OVIPOSITION
Pachydiplax females were infrequently seen at the three ponds.
During July and August only 25 matings were closely observed. Females appeared between 3oo and 5oo hours (E.S.T.) on warm sunny days and courtship commenced immediately. A territorial male, upon seeing a emale within his territory, would fly directly to a position above her. Hovering a ew centimeters above her he would raise his abdomen in a manner similar to that seen in the threat display. (But in a position that makes it difficult (or the female to see this display.) He then flaps his wings rapidly in an unusually wide arc. A receptive emale will permit the male to align his body directly above hers. The male then descends and the female rises up slightly while lowering her abdomen and presenting her head (Fig ,c). In rapid sequence he clasps her with his abdominal appendages and they go into copula. Copulation occurs in flight and is brief, lasting o-4o seconds. Its durati.on appears to depend, in part, on the number o previous, inseminations received by the female, but only by collecting data could this speculation be verified. The pair then separate and the male returns directly to his display perch. On three occasions the male was seen to transfer sperm to his genitalia during this short postmating flight. Typically, the female also rests for 5-o seconds on a nearby perch before searching for oviposition sites.
Oviposition was observed as illustrated by Needham and Wesffall (I955). I found that the male defended the female from interference by other males only as long as she remained within his territory. The flight of these deending males was more directed and intense than that seen in male-male interactions. Nonetheless, in a few cases, when many males were present and/or when the female strayed out of her mate'.s territory, his fervent defense was inadequate and oviposition prevented. A behavior, marked' in Libellula incesta Hagen but less common in P. longipennis, occurred when unmated males attempted to grasp the female while her head was exposed during oviposition. The female was usually knocked, tumbling, into the water. L. incesta females, being stronger fliers, would attempt to evade such attacks; P. longipennis females usually retreated to the trees until male excitement diminished. In addition to interference from other males, rogs presented a major threat to females ovipositing near the edge of the ponds.

INTERSPECIFIC BEHAVIOR
Pachydiklax longikennis males are unusually aggressive dragonflies when compared with other Libellulids. They react with partieular ferocity to three larger sympatric species whose pruinosity color is similar to the grey-blue color of Pachydillax longipennis.
The first, Libellula cyanae Fabricius, most common at Pickman and the Five Fields ponds has a slightly darker pruinescence, covering the entire adult male body. Libellula incesta, .common only at the Shade Street pond, has a. much darker, almost black, pruinescence. The third, Erythemis simllicicollis Say, has the same color as Pachydillax longipennis, covering a teneral green on the dorsal surface, of the abdomen and synthorax. It occurred at all three of these ponds.. Libellula incesta and L. cyanae sometimes occupy perches and enter the territory of Pachydiplax longikennis. Erythemis simllicicollis usually preferred to perch on an e.xpos.ed log or rock, situations rarely used by P. longilennis. Despite its smatler size, P. longilennis would display and chase all three of these species that entered its territory.
Other dragonflies commonly present at these ponds included: Plathemis lydia, 8ymetrum rubicundul.um, Leucorrhina intacta, Gomlhus furcifer, Perithemis tenera, Libellula ulchella, L. luctosa and Ladonia julia. P. longikennis males generally ignore them, presumably because of their different appearance. COLORATION Coloration patterns o. recently dead males (killed by chilling) were compared in visible and ultraviolet light. It was ound that the blue pruinose areas on the .a'bdomen though dull under visible light have marked reflectivity in the ultraviolet. Highly reflective pruinescent areas were also found in males of Libellula cyanae, L. incesta and Erythemis simlicicollis (see Fig. 2), but were absent in emales of these species Some odonates, including a libellulid, have been shown by physiological methods to be UV-sensitive (Goldsmith and Bernard, 974: Table 5). Why is it that these, more. stationary, libellulids have pruinescence in similar areas while in others it is a different color, located in different areas or absent all together? Species such .as Libellula kulchella have only white pruinescent spots on their wings and are much more mobile than P. longikennis. In light of these preliminary findings, the possible role of ultraviolet patterns in odonate behavior deserves careful attention.