EMERGENCE PATTERNS OF THE SUBALPINE DRAGONFLY SOAIA TOCHLORA SEMICIRCULARIS (ODONATA: CORDULIIDAE)

The genus Somatochlora is circumpolar and many o its species reach high into the subarctic and subalpine regions. S. semicircularis occurs in the Canadian and Hudsonian zones o the North American Western Cordillera rom Alaska to Colorado. Towards the southern part oi: its range, this species is ound at high altitudes in Colorado where adults have been recorded at 3000 to 3700 m (Gloyd, 939; Walker, 925; Carpenter, 873). Larval habitats range rom 2700 to 36o0 m in shallow ponds and water meadows. These may well be the highest breeding localities or any odonate known. The alpine and subalpine habitats, offer extremes in temperature and insolation during the short growing season which ollows the late snow melt (Mani, 968). The subalpine regions, particularly, present relatively the highest day and lowest night temperatures (Billings and Mooney, 968). Coupled with these actors, the requent periods o drought and general unpredictability o climate contribute to the rigor ot: the high altitude habitat. Studies o lie history strategies ot: insects which have adapted to high altitudes have revealed cold resistance and lowered temperature thresholds, a predominance o( diurnal activity, a seasonal delay and shortening o( the flight period, and a general increase in the length o the lie cycle coupled with overwinter hibernation or diapause (Mani, 968). In the dragonflies, the increased li(e cycle is accomplished by elongating the larval period to as much as three or four years (Paulson and Jenner, 97I Robert, I958). Somatochlora semicircular# larvae are also resistant to dessication when their ponds and water meadows dry up annually (Willey and Eiler, 972). An ecologically sensitive part o( the lie history or a dragonfly is the emergence period during which the "aquatic" larva leaves the water, trans(orms into the *Department o Biological Sciences, University of Illinois at Chicago Circle, Chicago, Illinois 60680; and The Rocky Mountain Biological Laboratory, Crested Butte, Colorado 81224. Manuscript received by the editor February 7, 1974.


INTRODUCTION
The genus Somatochlora is circumpolar and many o its species reach high into the subarctic and subalpine regions. S. semicircularis occurs in the Canadian and Hudsonian zones o the North American Western Cordillera rom Alaska to Colorado. Towards the southern part oi: its range, this species is ound at high altitudes in Colorado where adults have been recorded at 3000 to 3700 m (Gloyd,939; Walker, 925; Carpenter, 873). Larval habitats range rom 2700 to 36o0 m in shallow ponds and water meadows. These may well be the highest breeding localities or any odonate known.
The alpine and subalpine habitats, offer extremes in temperature and insolation during the short growing season which ollows the late snow melt (Mani,968). The subalpine regions, particularly, present relatively the highest day and lowest night temperatures (Billings and Mooney, 968). Coupled with these actors, the requent periods o drought and general unpredictability o climate contribute to the rigor ot: the high altitude habitat. Studies o lie history strategies ot: insects which have adapted to high altitudes have revealed cold resistance and lowered temperature thresholds, a predominance o( diurnal activity, a seasonal delay and shortening o( the flight period, and a general increase in the length o the lie cycle coupled with overwinter hibernation or diapause (Mani,968). In the dragonflies, the increased li(e cycle is accomplished by elongating the larval period to as much as three or four years (Paulson and Jenner, 97I Robert,I958). Somatochlora semicircular# larvae are also resistant to dessication when their ponds and water meadows dry up annually (Willey and Eiler,972 winged adult, and flies away as a reproductively immature (teneral) adult. The observations reported here are intended to correlate a late emergence season, a possible lowered temperature threshold, a complete shift to diurnal behavior, and an abbreviated, synchronized emergence period with the rigors of the high altitude habitat in Colorado.
HABITAT Two subalpine ponds were selected for study areas. They represent the typical habitat where larvae of Somatochlora semicircularis are found. The emergence areas of both ponds are restricted to the shallow, sedge-filled margins.
Pond Nine (346 m) is one of a group of small ponds in the Mexican Cut Research Area on Galena Mountain (Gunnison Co., Colorado) owned by The Nature Conservancy and administered by the Rocky Mountain Biological Laboratory. It is o m in diameter and m deep in the center. About 50% of its shoreline is lined with Carex rostrata and C. aquatilis in the shallow bays which make up the emergence areas for the dragonflies. The center of the pond is 1974] [Cilley 8ubalDine Dragonfly and shallow areas where emergence takes place are lined with C. rostrata and C. aquatilis. The deeper areas support Potamogeton foliosuv and P. illinoisensis and thin mats of Drepanocladus uncinatus. This pond has been described further by Willey and Eiler (1972).

MATERIALS AND METHODS
The emergence of Somatochlora semicircularis was studied from I968 to I972. At first, the transforming larvae were marked by surveyor's flags so that the transformation of each individual could be accurately timed. Considerable trampling of the habitat was unavoidable so that the studies were repeated in I97I by making daily collections of the cast skins left behind by the teneral adults. Fewer larvae were damaged during their transformation by this method.
However, more larvae were lost through bird predation so that their cast skins could not be recorded. In order to verify the diurnal pattern of the teneral flight in an undisturbed habitat, continual observations by binoculars of the ponds were conducted from OO feet away from the shoreline in I972 during the peak of the emergence period.
Water temperatures were measured 3 cm below the water surface in areas of highest emergence activity. Solar radiation was measured by a recording Belfort pyrheliograph with a 2-hour gear. Environmental air temperatures were measured with a YSI telethermometer shielded from radiation by thin, flat, double-layered shields, 46 cm square, which were set horizontally 8 cm above and below the temperature probe. The inner surface of each was painted flat black and the outer surface was glossy white. (Platt and Griffiths, I964).  away from the ponds as teneral adults (maiden flight). At Pond Nine, emergence began on I5 July I968, 7 July 1969, and 6 July 97o. At Irwin Pond South, slightly lower in altitude and more heltered by spruce-fir, emergence began slightly earlier--30 June 1970, 3 July I97 I, and 27 June I972.
The total period during which emergence occurs each year is short (Table ). The total annual production of adults occurs over an average of I6 days during which 50% of the adults emerge within the first three to six days. The I969 emergence curve (Fig. 2) demonstrates a typical, highly synchronized pattern. The 1968 curve ( Fig. 3) shows the effect of stormy weather and low temperatures on I7 and 8 July which delayed the emergence of some of the larvae and extended the emergence period. No secondary peak of emergence was ever observed later in either season. The distance flown during the initial flight of the teneral adults is unknown. However, they remain away from the ponds, presumably in the forest near the ponds or at lower altitudes, for one to two weeks. In I968, the first tenerals left Pond Nine on 15 July and the first mature male's were observed patrolling the edges of the pond on 22 July 7 days later. By this time, 95% of the adults had emerged and flown away (Fig. 3). In 969, the first tenerals left on 7 July and the first patrolling male was sighted on 22 July--2 days after the last teneral adult had flown (Fig. 2). These mature adults are individuals which emerged at the beginning of the' emergence period, have matured, and returned to the pond for mating. Teneral adults and reproductively mature adults are considered here to be all part of the same emerging population. They are just temporally separated. On two occasions mature males, were observed patrolling a pond when a newly emerged teneral flew up from the sedges on its maiden flight. On both occasions, the mature male harassed and tried to couple with the soft te'neral befo.re the latter could leave the pond. One of the tenerals was probably injured by 1974] tl/illey 8ubalpine Dragonfly 125 Psyche [March the encounter because its subsequent flight was very erratic. The 8omatochlora males are very aggressive and will try to grasp any other dragonfly--its own species, Cordulia shurtle, or even the much larger Aeschna interrupta. Water and air temperatures are relatively low in ponds at high altitudes although the diurnal water temperatures are not as low as might be expected due to the heat absorbing qualities of the pond bottoms and the shallowness of the water. The water temperature averages I2C. (varies from 9 to 4C.) when the larvae trst emerge in the morning, whereas the air temperatures, which rise very rapidly once the sun is up, average 9C. (vary rom 4 to I2C.).
The night air temperatures drop below reezing and, on many mornings, frost still lay on the ground when the iqrst larvae appeared. The larvae climb up the sedges in areas where the water is seldom more than 30 cm deep and usually 3 to 8 cm deep. The larvae start to leave the water after sunrise (0700 to 0800 MDT).
The maximum number of larvae leave the water between 0800 and Ioo hours (Fig. 4). By ooo to 7oo, they are ready to fly.
In 972, observations of undisturbed ponds indicate that 75% or more of the tenerals actually leave the ponds between oo tnd 23o i the weather is fair and the air temperature varies between 6 and SC. The initiation of the flights appears to be temperaturedependent. On July 972, flights began at IOO with an air temperature o 8C. The flights stopped abruptly at 23o when clouds covered the sun and the air temperature dropped to 4C. Flights did not resume until 43o when the sunshine returned and the air temperature rose to 6C. All flights ended for the day at I5oo when a thunderstorm started and the air temperature dropped to I4C. and below. Any adults which had not flown by this time had to remain on the sedges until the o.llowing morning.
Birds are the main predators o Somatochlora during the actual transformation process. Spiders and ants will attack larvae near the shore but generally account for a very small proportion of the loss. Speci(ic data are being assembled for a separate report on bird predation and will not be presented here. However, a general summary is essential for the present argument.   Adults which have trans(ormed but have been prevented rom flying by low temperatures or rainy weather may remain clinging to the sedges or a day or more. Most o these adults all prey to birds. Once the tenerals fly, they are reasonably ree rom bird predation. Only one teneral has been observed to be caught by a bird (Western flycatcher) during its maiden flight.

DISCUSSION
Somatochlora semicircularis is a western cordilleran dragonfly whose adults normally appear between the middle o June and the middle oi: July (Walker, 1925). The Colorado ponds under investigation occur at the southern limit o( the species and yet the adults emerge in early July and appear as reproductives in mid-July near the end o the general emergence season. Such a relatively late season is probably correlated with the late snow melt and low temperatures o its high altitude habitat. Schiemenz (1952) has reported a similar delay o( season correlated with altitude or Libellula quadrimaculata. Dragonfly emergence has been shown to be temperature-dependent with the initiation o emergence primarily dependent on water temperature (Trottier, 1973;Corbet, 1963). The larvae o S. semicircularis in Colorado first leave the water when the water temperatures average I2C. and the air temperatures average 9C.
Comparative data or other corduliid dragonflies are not available. The only data available come rom the studies o the aeschnid dragonfly d nax. Corbet (1957)  At water temperatures below I7C., the .4nax larvae which had let the water would return to the water and delay emergence until the ollowing day. The Somatochlora larvae were able to leave the water and undergo trans(ormation at average temperatures 5C. below the minimum or .4nax larvae. Whether this difference is a specialization to high altitudes or is only a general amily difference must be urther investigated. In general, teneral corduliid dragonflies at lower altitudes and/or latitudes are ready to fly early in the morning. However, the emer- Hours of the Day Figure 4. Daily emergence pattern of 8. semicircularis on 2 July 1970 at Irwin Pond South. Lower curve shows the number of nymphs leaving the water (solid line) and transformed teneral adults ready to fly (dashed line) at successive hours during the day. Middle curve shows water temperature cm below the surface at two separate stations, m apart, and air temperature 45 cm above the surface of the wa,ter in the emergence area. Top curve shows solar radiation. Nymphs recorded he,re completed the entire transformation process successfully.

I3o
Psyche [March gence pattern is variable and reshly emerged adults have been collected throughout the day and even in early evening on cold, cloudy, or rainy days (Lutz and McMahan,973;Corbet,963;Lutz,I96 Taketo,96o;Kormondy,959;Robert,958 Lietinck,93:)-The larvae of 8. semicircular#, under normal conditions, do not leave the water until after sunrise and the tenera.ls generally fly at noon. A similar shift of the emergence pattern to the middle of the day has been described for J eschna interrupta at high altitudes in California (Kennedy, I925) and for Jeschna subarctica on the moors near Kiel, Germany (Schmidt,968).
The emergence curve of Somatochlora in Colorado is characterized by brevity (6 days) and a high degree of synchrony (Fig. 2, Table ). Similar synchrony has been described for other dragonfly species which have been categorized by Corbet (963) as "spring" species. Only one of these species, Oldonaeschna armata, exhibits a shorter time for the total period (Io days) and for the 5o% emergence time (2 days). Oplonaeschna inhabits the water-pocket streams of New Mexico and Arizona and is subject to flash floods and severe dessication (Johnson,968). The lower altitude corduliid, Tetragoneuria cynosura, averages a 23-25 day emergence period with 5o% of the adults emerging by the 6th or 7th day in Michigan (Kormondy, 959) and North Carolina (Lutz and McMahan,973 Lutz, 96 and 962). Tetragoneuria in Michigan also exhibits a bimodal curve in which a second peak of emergence occurs o days after the end of the irst peak. This bimodality has been described for other dragonfly species (see Corbet,963) but is missing in the 8omatochlora curve (Figs. 2 and 3).
Selective pressures contributing to the synchrony of the single peak of the Somatochlora emergence curve may well involve the brevity of the season. Both in 968 and 969, o days a(ter the end o the initial emergence peak at a time when T. cynosura (Kormondy,959) and ACnax imperator (Corbet,957) normally exhibit a small second peak o emergence, the shallow areas at Irwin Pond South and Pond Nine had largely dried up. 8omatochlora larvae leave the water by climbing up the emergent sedges. They do not crawl out on the land. Lack of emergence sites or competition or the ew remaining sites can be expected to be a actor in selecting or a temporal shift toward early emergence.
Another contributing factor may involve intrspeci(ic interference. The diurnal emergence pattern which is necessitated by the climatic character of the habitat creates problems which normally do not affect 1974] Willey 8ubalpine Dragonfly nocturnal or early morning emerging populations. Teneral adults which fly up during the middle o the day are subject to harassment and possible damage rom the reproductive males. However, by the time the first 8omatochlora reproductives return to the Colorado ponds, 95 to oo% o. the annual production o( tenerals have already lei:t the pond (Figs. 2 and 3). The diurnally temporal separation o.i: the reprodu.ctives and the nocturnal or early morning tenerals o the lower altitudes has been replaced by the seasonal separation o a to 2 week teneral maturation period during which time the dragonflies fly in the orest away :rom the ponds while the remainder o the population is emerging. The high altitude ponds o( Colorado seem to be suciently isolated rom each other to keep recruitment o( reproductives rom lower, and therefore earlier, ponds to a minimum. The diurnally transforming dragonflies are also exposed to predation by birds. The highly synchronized emergence o 8omatochlora, however, provides that the relatively large number o larvae which emerge during the rst ew days ensure that many tenerals get away :rom the pond. To what extent a synchrony o. emergence provides selective advantage to the dragonfly population by () partially orestalling the results o a revised predator searching image and/or shift in prey preference or (2) saturating the environment with prey with a resulting early satiation o the predators is as yet undetermined. It is likely that both predation tactics are involved and thwarted by this mechanism. UMMARY Somatochlora semicircularis has been able to adapt to the rigors o the subalpine and lower alpine habitats o Colorado. The corduliid tendency toward a flexible, early morning emergence pattern has contributed to its shit to a Cull diurnal emergence pattern. In addition, it may be able to initiate emergence at relatively low water and air temperatures. The emergence pattern has been strongly synchronized, shortened and restricted to a single seasonal peak, possibly as a result o ponds drying up early in the season at high .altitudes, predation pressure, which may be the same or worse than that at lower altitudes, and the effects o potential interference between the aggressive reproductive males and the sot-cuticled teneral adults. JixCKNOWLEDGEMENTS I wish to thank Dr. Robert B. Willey or constant advice and encouragement. I appreciate the valuable critical reviews by Drs.