Introduction
Neoplasta parahebes MacDonald and Turner (Diptera: Empididae: Hemerodromiinae), a small predaceous dance fly, occurs in western North America from southern British Columbia to southern California. The taxonomy of European Hemerodromiinae, based on adult morphology, is well known, but knowledge of the life history is incomplete (Wagner Reference Wagner and Nilsson1997). A similar situation exists in North America.
Species of Neoplasta Coquillett are found throughout the New World from Patagonia, Chile, and Brazil (Collin Reference Collin1933; Smith Reference Wagner and Nilsson1962) to the United States of America and Canada (Melander Reference Melander1927, Reference Sinclair and Harkrider1947). Twelve species occur in North America, north of Mexico (MacDonald and Turner Reference McAlpine, McAlpine, Peterson, Shewell, Teskey, Vockeroth and Wood1993). MacDonald and Harkrider (Reference MacDonald and Turner1999) used DNA analysis to associate adults and larvae of Neoplasta and differentiate the latter from larval Metachela Coquillett (Hemerodromiinae). Brammer et al. (Reference Brammer, Harkrider and MacDonald2009) used the same technique to differentiate the immature stages of all four genera of Hemerodromiinae found in North America.
In spite of their wide distribution and occasional local abundance, the biology of Neoplasta species remains largely unknown. Adult emergence studies in Quebec indicated a preference by N. scapularis (Loew) for swift water (Harper Reference Landry and Harper1980). Landry and Harper (Reference MacDonald and Harkrider1985) showed that immature N. scapularis are found in small and large streams in Quebec and that adults emerge over a 10 week period. Similar patterns were reported by MacDonald and Turner (Reference McAlpine, McAlpine, Peterson, Shewell, Teskey, Vockeroth and Wood1993).
Harkrider (Reference Harkrider2000a) found that 78% of adult Neoplasta collected along San Antonio Creek, a California mountain stream, were N. parahebes. In the only detailed study of immature aquatic empidids in North America, Harkrider (Reference Harper2000b) observed that larval Neoplasta are predators of larval Rheotanytarsus Bause (Diptera: Chironomidae) in that stream. The objective of the present study was to document the life history of N. parahebes at the California site.
Materials and methods
Study sites
The primary study site was a section of San Antonio Creek (“creek”) approximately 300–350 m from the Mount Baldy Road bridge northeast of Baldy Village in the San Gabriel Mountains of southern California (Fig. 1; 1250 m, 34°14′25″N, 117°39′10″W). The surrounding vegetation was dominated by white alder (Alnus rhombifolia Nutt. (Betulaceae)) and California laurel trees (Umbellularia californica (Hook. and Arn.) Nutt. (Lauraceae)). This site had water year-round, although the creek is often dry above and below the primary study site during the summer. A second study site (“small stream”) was a first-order stream fed from the spring, running year-round in alluvial deposits paralleling the west bank of San Antonio Creek (Fig. 2; Harkrider Reference Harkrider2000a).
A third site (“Thurman Flats”) was a seepage area in the San Bernadino Mountains along Highway 38, about 25 km east of San Bernardino (approximately 1030 m; 34°6′11″N, 117°0′46″W). The surrounding vegetation was dominated by white alder and California blackberry (Rubus ursinus Cham. and Schltdl. (Rosaceae)).
Specimen collection and species separation
Adult N. parahebes were collected from stream-side vegetation at the creek and small stream sites with a sweep net and stored in four dram vials. Immature stages were collected by hand from rotting wood, midge tubes, vegetation, and other materials in the water. For purposes of morphological comparison and species separation, adult Neoplasta hebes (Melander) were collected by sweeping vegetation at the Thurman Flats site. All specimens were stored on ice for return to the laboratory.
Life history
Field-collected gravid female of N. parahebes were dissected under a microscope using No. 3 insect pins for the fecundity studies. Oviposition units consisted of a 450 mL base with a clear plastic 450 mL top as described in Sinclair and Harkrider (Reference Smith2004). Various objects, such as wet paper toweling, branches, or vegetation, were placed in the units to stimulate oviposition. Eggs were moved to Stentor dishes for incubation. Larval rearing units were 250 mL, open-top containers as described in Harkrider (Reference Harper2000b). Larvae of a Rheotanytarsus species or Orthocladius lignicola Kieffer from the creek collecting site were provided as prey.
Larval body, head complex (the sclerotized posterior tip of metacephalic rod to end of labrum), and eggs were measured using an ocular micrometer or by measuring digital images made with a dissecting or compound microscope and Motic Images 2000 (Windows XP version 1.2, 2000) software. Head capsule lengths of larval N. scapularis and N. parahebes were compared using a two sample t test.
In 1999 and 2000, emergent branches, 10–20 mm diameter and showing some evidence of deterioration, were collected from the small stream site near a resting site for adult Neoplasta. In the laboratory, the branches were systematically searched for eggs. Eggs were removed using dissection needles and moved to Stentor dishes for rearing.
In 2002, branches in contact with the cobble substrate, 5–28 mm diameter and showing signs of decay, were collected from the creek and small stream sites at 2-week intervals from January through September. In the laboratory, bark was removed and soft tissue was probed with dissection needles under a dissecting microscope. Presence of midge tunnels, midge larvae, empidid larvae, and other insects was recorded. In 2003, similar branch collections were made from the small stream site. In addition to the dissection of soft tissue, the branches were maintained in water without aeration for 24 h. The resulting reduced oxygen tension caused many additional larval Neoplasta to emerge from midge tunnels in the wood.
Surface area of field-collected branches was estimated by calculating the surface area of a cylinder using branch diameter (measured at both ends) and length measurements. In all studies, live larvae were placed in rearing units. Damaged or surplus larvae were killed with hot water and preserved in 70% ethyl alcohol.
Results and discussion
Adults
Species separation, phenology, and abundance
Harkrider (Reference Harkrider2000a) was unable to differentiate females of the two species of Neoplasta occurring at San Antonio Creek because diagnostic characters had not been described and the females “appear inseparable” (MacDonald and Turner Reference McAlpine, McAlpine, Peterson, Shewell, Teskey, Vockeroth and Wood1993). Comparison of the dominant female type from this site with females collected during this study from Thurman Flats (where only N. hebes males were found) revealed that the ovipositor of N. parahebes is thinner at the proximal end and possesses a ventral keel-like structure (Fig. 3B) that is lacking in the ovipositor of N. hebes. Using the ovipositor diagnostic character, Harkrider's (Reference Harkrider2000a) original estimate of the relative abundance of N. parahebes at San Antonio Creek has been slightly revised to 75% of aquatic empidids.
Some adult male N. parahebes were found as early as March, and a few individuals remained until December. The greatest abundance of adults occurred from May to July; male abundance peaked in early May, and female abundance peaked in July.
Mating
Mating pairs of N. parahebes were observed twice. A pair was observed in copula on the leaves of a California laurel tree above the small stream on a June morning. A second pair was observed in copula in a rearing unit after having been captured that day. In both cases, the mated pairs were oriented unidirectionally with the male above the female as described by McAlpine (Reference Melander and Wytsman1981). The female of the captured pair stood on a ledge of the rearing unit with the male grasping her, his head above her thorax. The female shook violently with a yawing motion for 2–3 s, and then stopped for 4–10 s before shaking again. This behaviour continued for about 3 min and was subsequently repeated, though shorter in duration, four times during the next 36 min. All movements were abrupt and the pair was usually motionless between movements (the female occasionally turned or moved short distances). After the pair broke apart, they oriented face to face, less than a body length apart, for a second or two. The male then flew away and no further interaction was observed.
Feeding
Adult N. parahebes were observed in 8 or 16 dram vials with conspecifics, adult N. scapularis, and various potential prey items collected from the same habitat. Cannibalism was never observed between adults of N. parahebes, but in one instance, a male N. scapularis fed on a male N. parahebes. In all instances where feeding by adult N. parahebes was observed, the prey were midges that were smaller than the adult empidid.
Oviposition
In June and July of 1999 and 2000, 29 branches were collected. Of these, 11 had eggs deposited in cracks or under loose bark; the eggs were not visible without removing the covering bark or debris. Egg density ranged from 1.3 to 47.8 eggs/cm2 of the branch surface. Eggs were deposited in small clusters or individually over a wide area. Eggs did not occur under bark still firmly attached to the wood beneath. Most eggs were deposited on submerged sections of the branches.
Initial attempts to induce oviposition of gravid females in rearing units failed. Most potential oviposition substrates, as well as strategies such as decapitation of gravid females (Cummings and Cooper Reference Harkrider1993), failed to induce oviposition. However, decaying branches from the stream stimulated oviposition and resulted in egg deposits similar to those on field collected branches.
In the laboratory, a female of N. parahebes laid most of her eggs underwater within a 24 h period, although actual oviposition was not observed. Dissection of 10 gravid females revealed a range of 62–88 mature ovarioles (average 74.8). Dissected females in early stages of egg development showed as many as 96 ovarioles.
Immatures
Eggs
Length of 30 eggs laid by three females in a rearing unit ranged from 418 to 502 μm (average 459 μm). Width ranged from 106 to 124 μm (average 115 μm). Eggs collected in rearing units and held in Stentor dishes produced larvae within 9–11 days at room temperature.
Larval development
Length of newly hatched N. parahebes larvae ranged from 600 to 800 μm (720 ± 90 μm, n = 10); length of head complex ranged from 160 to 170 μm (163 ± 3 μm, n = 10). The newly hatched larvae would not feed and all died within 2 or 3 days. Thus, cannibalistic behaviour common in later stages was not observed.
Harkrider (Reference Harper2000b) suggested that the head complex of the last instar larva of N. scapularis was significantly larger than that of N. parahebes, but this was based on only one specimen of N. scapularis. In the current study, lengths of the head complex of last instar exuviae from reared specimens of N. scapularis (580 ± 20 μm, n = 10) and N. parahebes (545 ± 28 μm, n = 10) were significantly different (t = 3.63, P = 00.0019, df = 18).
There appear to be three larval instars in N. parahebes. Harkrider (Reference Harper2000b) noted two head complex size classes in larval Neoplasta (0.50 mm and 0.30 mm). In the current study I plotted body length versus head complex length of larval N. parahebes collected in three different years (Fig. 4). Some larvae longer than 4 mm were collected but were used for rearing and were not measured. Although it is not possible to identify larval Neoplasta to species and thus, more than one species may be represented in the data, three distinct head length groups are evident: 0.16–0.18 mm (n = 3), 0.30–0.35 mm (n = 15), and 0.50–0.60 mm (n = 24).
Cummings and Cooper (Reference Harkrider1993) reported three larval instars for tachydromiine empidids and Sinclair and Harkrider (Reference Smith2004) suspected the presence of three larval instars for Roederiodes wirthi Chillcott (Clinocerinae). A larval Neoplasta field collected in July 2004 had a head complex measuring 0.16 mm; it molted 2 days later to a larva with a head complex of 0.30 mm. These measurements agreed with the first two size classes (instars) suggested in Figure 4. The data reported here support the hypothesis that three larval instars occur in Neoplasta.
Larval habitat
Harkrider (Reference Harper2000b) recorded larval Neoplasta in larval tubes of Rheotanytarsus midges, but a search for other larval habitats was not successful. In 1999, eggs, larvae, and pupae of Neoplasta were found on emergent branches. From January through September 2002, larval densities were relatively similar on small stream and creek decaying branches (Fig. 5). Densities were highest at the start of the year, and from February through May, average density ranged from 1.8 to 2.5 larvae/100 cm2 of branch surface area. A clear reduction in larval densities was observed in June, and from July to September, densities declined from 1.1 to 0 larvae/100 cm2 of branch surface area. Reared adults indicated that N. parahebes and N. scapularis occurred at both sites.
During the 2002 survey, larval Neoplasta were strongly associated with larvae of the wood-boring midge Orthocladius lignicola. Only one larval Neoplasta was found in wood without evidence of O. lignicola. Larval O. lignicola burrow several millimeters into decaying wood, creating a pattern of tunnels in which larval Neoplasta can be found.
In 2003, 41% of larval Neoplasta collected in January came from samples of decaying branches held in water for 24 h without aeration after dissection to induce the larval empidids to vacate the midge tunnels. A large number of these larvae were small and would have been missed during the 2002 survey. The January maintenance of branches for 24 h without aeration yielded a total of 25 Neoplasta larvae (34% of all larvae collected in 2003). The January 2003 small stream branch samples produced 9.2 larvae/100 cm2 (Fig. 5), nearly three times the density observed in January 2002. From January to April 2003, density of larval Neoplasta declined in the small stream branch samples, which is similar to the pattern observed in 2002.
The pattern of decline of the populations of larval Neoplasta is consistent with increasing numbers of pupae and consequent emergence of adult Neoplasta from May through July as observed by Harkrider (Reference Harkrider2000a). Cannibalism or migration of larvae to other habitats such as larval tubes of Rheotanytarsus midges could also be a factor.
Although 22 of 26 branches sampled in the first 3 months of 2003 contained larval O. lignicola, only 14 of the branches also contained larval Neoplasta.
Pupal habitat
Harkrider (Reference Harper2000b) found no pupal Neoplasta in 4234 larval tubes of Rheotanytarsus midges. Five pupal Neoplasta were found in the 2002 survey of branches and 12 pupae were found within O. lignicola tunnels in the 2003 survey. The majority of adults reared from the 2002 and 2003 pupae were N. parahebes (n = 12); the remainder were N. scapularis (n = 4) and N. hebes (n = 1). Most pupal N. parahebes were collected from March to May. Pupal collection and adult rearing data are consistent with adult emergence and species relative abundance reported by Harkrider (Reference Harkrider2000a). Harkrider (Reference Harkrider2000a) also suggested that the observed long flight period of N. parahebes was due to the long survival of adults. However, the collection of a single pupal N. parahebes in August 2003 suggests that an extended emergence period is more likely.
Acknowledgments
I thank J.F. MacDonald (Purdue University, West Lafayette) who kindly reviewed an early version of the manuscript, Hans Meyer (Ökologie-Zentrum Christian-Albrechts-Universität zu Kiel) whose review was particularly helpful for data presentation, and Brad Sinclair (Canadian Food Inspection Agency, Ottawa, Ontario) whose review comments and encouragement were especially important.
