Prey selection and diurnal activity of Holcocephala oculata (F.) (Diptera: Asilidae) in Costa Rica

9 Pages·2000·4.3 MB·

by  L M Lapierre

PROC. ENTOMOL. SOC. WASH.
102(3), 2000, pp. 643-651
PREY SELECTION AND DIURNAL ACTIVITY OF
HOLCOCEPHALA OCULATA (F.) (DIPTERA: ASILIDAE) IN
COSTA RICA
Louis M. LaPierre
Department of Organismic Biology, Ecology, and Evolution, University of California,
Los Angeles, CA 90095-1606, U.S.A. (e-mail: [email protected])
—
Abstract. Holcocephala oculata (F.) is an abundant small robber fly in open areas at
La Selva Biological Station, Costa Rica, and was studied during August 1995. The ma-
jority (92%) of prey taken by H. oculata belong to the orders Hymenoptera, Diptera, and
Coleoptera, with the remaining belonging, in order ofdecreasing frequency, to Hemiptera,
Thysanoptera, Homoptera, Araneida, Strepsiptera, and Psocoptera. A comparison ofsticky
trap samples offlying insects and robber fly prey items reveal differences in theproportion
of orders represented. Either H. oculata prefers particular orders of prey, or sticky traps
are inadequate in sampling insect faunas. Cannibalism was not observed for H. oculata.
Flies are more numerous in sunny areas at the beginning and end of the day. Flies are
rarely observed in exposed areas on clear days when incident radiation is high. Areas
shaded from direct incident radiation through the day show little change in fly numbers.
Courtship and copulation, described herein, generally occur after 1200. Holcocephala
oculata perches less than a meter off the ground, on average, with grass species tending
to be the substrate most often chosen. No evidence forinvertebrate predation onH. oculata
was found, although several suitable predaceous arthropods co-occur with the robber fly
and at least one, the ponerine ant Ectatomma ruidum Roger, readily attacked robber flies
during feeding trials.
Key Words: Asilidae, Holcocephala oculata, robberfly, prey selection, courtship, diurnal
activity. La Selva Biological Station, Costa Rica
Robber flies are among the larger and item externally. In addition, some members
more visible members ofthe Dipterabut are of this group have courtship displays that
poorly understood when it comes to details usually involve characteristic aerial maneu-
of their ecology and behavior. A number of vers on the part of the male (Fisher and
temperate zone Asilidae have been studied Hespenheide 1982, 1992).
in detail (Lavigne and Holland 1969; Hes- Holcocephala oculata (F) (Dasypogoni-
penheide and Rubke 1977; Hespenheide nae: Danilini; Fig. 1) is the smallest (mean
1978, 1989; Weeks and Hespenheide 1985), length 5.20 mm; n = 10) of four species of
but only a few tropical species (Fisher Holcocephala that are known to occur at La
1983; Shelly 1984a, 1988; Fisher and Hes- Selva Biological Station, Costa Rica (E.
penheide 1992). These flies typically cap- Fisher and H. A. Hespenheide, unpub-
ture their prey by waiting at a perch, inter- lished). Adults perch on low-growing veg-
cepting the prey in mid-flight, and then re- etation in open areas. Here I describe the
turning to the perch and consuming the prey selection and diurnal activity of H.
644 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
("shady"). In addition to the transects,
neighboring weedy margins within the
Huertos project were used for additional be-
havioral and ecological observations on H.
oculata. The weedy margins are cut at three
to six month intervals and maintained at a
height ofless than three meters, and include
several species of grasses, the most com-
mon being Paspalum conjugatum Berg.,
Digitaria sp., and Cynodon dactylon (L.)
Pers. In addition, sedges (Cyperus spp.),
and a diverse assemblage of plant species
adapted to disturbed areas are common
within the margins, as are fallen stems and
leaves from adjacent tree plots.
At least five additional asilid species
(three Holcocephala, one Atractia, one
Mallophora) also co-occur with H. oculata
in the weedy margins, and many species of
predaceous arthropods may be found there
as well (personal observation). The biology
of one other species (H. ajfinis (Bellardi))
has been studied at La Selva and is gener-
ally similar to that of H. oculata (H. A.
Fig. \. HoUocephala oculata showing typical Hespenheide, unpublished).
perch behavior.
To determine the diversity of prey taken
by H. oculata, I captured feeding robber
oculata and test the hypothesis that net in- flies in the sunny transect during the 10
cident radiation affects robber fly densities days and collected their prey items. Prey
more than ambient temperature or time of were collected by placing a plastic vial over
day. a robber fly disturbing it so that it released
the prey item into the vial, after which the
Materials and Methods
robber fly was released. Preliminary cap-
The data were collected during 10 sepa- tures revealed no detrimental effects due to
rate days between 9-29 August 1995 at the this method ofremoving prey from the rob-
Organization for Tropical Studies' La Selva ber flies; released individuals remained in
Biological Station which is located at the the transect and were often seen with a new
confluence of the Rio Puerto Viejo and Rio prey item moments later. This method of
Sarapiqui, Heredia Province, Costa Rica prey removal has been used by others (see
(McDade et al. 1994). Hespenheide and Rubke 1977; Hespenheide
I chose two 50 m transects along weedy 1978, 1989; Shelley 1984a, 1988). I record-
margins bordering mono- and polyculture ed the time at which the prey item was col-
plots within the "Huertos" project, an on- lected in order to observe any temporal pat-
going study on the sustainability of soil fer- tern in prey choice. Prey were not sampled
tility in reconstructed tropical ecosystems from robber flies in the shady transect.
(J. Ewel, personal communication). One To gather information on the diversity of
transect was continually exposed to the sun available prey items, I placed two sticky
from dawn until dusk ("sunny") whereas traps (21.5 X 28 cm transparencies coated
m
the other was shaded most of the time with Tanglefoot®) on poles 1 above the
VOLUME NUMBER
102, 3 645
ground in the sunny transect during study was measured at a single fixed location in
days 8-10 while the robber flies were being the shade between the two transects.
sampled. Only arthropods measuring up to To describe the courtship behavior of H.
4.3 mm (the maximum prey size collected oculata, I surveyed the transects during the
from H. ociilata) were considered in the study periods and observed courting pairs.
analysis ofthe sticky traps. I identified prey I recorded the time and noted whether the
and sticky trap samples to order. To deter- outcome of the courting attempt was suc-
mine ifH. oculata exhibits a preference for cessful or unsuccessful in initiating copu-
certain prey over others or is behaving op- lation.
portunistically, 1 performed two analyses to To determine whether or not H. oculata
compare the actual robber fly prey with the exhibits a preference for type or height of
available prey gathered from sticky traps. perch, I walked the weedy margins in and
In one analysis I compared a subsample of around the transects and recorded the
the actual prey collected during the days the choice of perch substrate, location on sub-
sticky trap samples were being employed strate, and height ofperch above ground for
(days 8-10). In a second analysis I included all H. oculata encountered. The location
the entire actual prey sample collected dur- was defined as precisely where on the sub-
ing the 10 sampling days. strate the fly chose to perch. Above ground
To determine the daily cycle of H. ocu-
height referred to the shortest distance be-
lata, I censused flies on all 10 days dividing tween the fly's position on the substrate and
each into five two-hour periods beginning
the ground.
at 0800 h and ending at 1600 h. I conducted For all analyses, data on mean values are
a census at the beginning ofeach ofthe five evaluated with the non-parametric Wilcox-
periods. Flies in the shady transect were
on sign rank test, and contingency tables
simply counted while those in the sunny
employing the G-test or Fisher's exact test
transect were counted and assigned to one
are used to compare percentile data.
of four activities: Feeding, copulating,
The arthropod predators most often ob-
courting, or other. The activity "other" in-
served in the transects foraging on sub-
cluded flies that were perched but not feed-
strates used by H. oculata were Ectatomma
ing and those that were flying between
ruidum Roger, an aggressive ponerine ant,
perches. During the intervening time peri-
ods, I kept out of the transects to minimize and orb-weaving spiders (Araneidae). I pre-
my formed feeding trials to determine the pal-
disturbing the flies.
Prior observations on the daily cycle of atability ofthis fly to E. ruidum with robber
H. oculata indicated that individuals avoid- flies disabled by having one wing removed.
ed areas exposed to direct sunlight during Orb-weavers construct webs in the weedy
the hottest times of the day, especially on margins at the level of H. oculata's perch
clear days. I gathered data on ambient tem- heights. Because prey items accumulate in
perature and incident radiation to determine these webs, information on the spider's prey
possible influences on the daily cycle ofH. choice is readily available. I carefully re-
oculata. My hypothesis was that net radia- moved the silk from entrapped prey to es-
tion was more important than temperature tablish whether H. oculata was among
or time of day in determining robber fly those preyed upon by the spiders. Other ar-
density in a given area; the null hypothesis thropod predators observed includedjump-
was that there was no difference among the ing spiders (Salticidae), which have been
three independent variables in determining observed to take robber flies (H. A. Hes-
robber fly density. Net radiation was mea- penheide personal communication), and ti-
sured with a net radiometer monitored by ger beetles (Cicindelidae). However, I ob-
the Huertos project. Ambient temperature served so few individuals (n = 4 and n =
646 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
2, respectively) that I am unable to evaluate (Lavigne and Holland 1969, Hespenheide
their potential as robber fly predators. and Rubke 1977). For H. oculata, compar-
isons of available prey from sticky traps
Results and Discussion
with actual prey collected during the same
I collected nine orders of arthropod prey period (days 8-10) and actual prey from the
from H. oculata (Table 1). Among these the remaining days (days 1-7) show dissimilar
Hymenoptera, Diptera, and Coleoptera to- proportions with respect to some prey or-
gether account for 92% of the total prey ders (Fig. 2). For days 8-10, Hymenoptera,
sample. The most frequent prey at the study Diptera, and Coleoptera are the predomi-
site were ant reproductives (28.5%). The nant orders of actual prey of H. oculata as
average prey size was 1.48 mm (n = 379) well as of sticky trap samples at 90% and
with a range of0.4 nmi (nematocerous Dip- 83%, respectively. In general, there is no
mm
tera) to 4.3 (Coleoptera). Hymenoptera, significant difference between the two sub-
Diptera, and Coleoptera were also found to samples of actual prey. This could indicate
dominate prey samples of Holcocephala that H. oculata is consistent in prey choice,
fusca Bromley and H. abdominalis (Say) in or that the relative availability of prey
the eastern United States (Dennis 1979, changed little over the course of the study.
Scarbrough 1982). However, H. calva The higher proportion of formicid Hyme-
(Loew), also from the eastern United States noptera among prey could be the result of
and while showing a similar preference for robber flies preferring this group. The high-
Hymenoptera and Diptera, differs from oth- er proportion of non-nematocerous Diptera
er Holcocephala studied to date by prefer- in the sticky trap samples could be due to
ing Hemiptera, Homoptera, and Psocoptera robber flies not favoring this group, perhaps
over Coleoptera (Scarbrough 1982). The because they are generally too difficult to
predator to prey size ratio for H. oculata is capture. These results are in contrast to
3.51 (mean robber fly length 5.20 mm). those found by Shelly (1984a) for Atractia
This is identical to that determined for H. marginata Osten Sacken in Panama where
fusca (3.5, mean robber fly length 5.9 mm; non-nematocerous Diptera composed a
Dennis 1979), and intermediate between H. lower proportion of prey relative to what
abdominalis and H. calva (3.4 and 3.7, re- was available from sticky traps, and Hy-
spectively, mean robber fly lengths 5.81 menoptera were insufficiently represented
mm and 7.25 mm, respectively; Scarbrough in either sample. Alternatively, the differ-
1982). Compared to other asilid genera, the ences observed between robber fly prey and
predator to prey size ratio for H. oculata is sticky trap samples may be an artifact of
close to that reported for the similarly sized the limited ability of sticky traps to ade-
robber fly Nannocyrtopogon neoculatus quately sample them, and of sticky traps
Wilcox and Martin (3.66, mean robber fly themselves being more attractive to some
length 6.62 mm) in the southwestern United taxa than others (Shelly 1984b, 1988; Men-
States (Hespenheidel978), but differs from sah 1996).
that reported for the much larger Sticho- Size comparisons among and between
pogon trifasciatus (Say) (2.23, mean robber actual and available prey also reveal signif-
fly length 13.00 mm), also from the south- icant differences (Table 1). In general, there
western United States (Weeks and Hespen- is no difference (P > .8; Wilcoxon rank
heide 1985). No incidence of cannibalism sum test) between the average size prey tak-
by H. oculata was observed. en by H. oculata and the average size prey
Factors that contribute to prey choice for available to it. Differences emerge, how-
particular asilid species likely include the ever, when comparisons are made between
relative difficulty in handling prey as well individual prey orders. These results indi-
as the availability ofprey in space and time cate that H. oculata prefers significantly
VOLUME NUMBER
102, 3 647
0.3-, n Actual prey (days 8-10)
Availableprey (days 8-10)
Z2 Actual prey (days 1-7)
Hym-F Hym-NF Dip-N Dip-NN
0.75-
o
c
o
o
&,
o
0.25-
0800 1000 1200 1400 1600
Time (hours)
Figs. 2-3. 2, Proportions of actual and potential prey of H. oculata: available prey from sticky traps (days
8-10; closed bars), actual prey from H. oculata (days 8-10: open bars), and remaining actual prey (days 1-7;
cross-hatched bars). Relationships among taxa that are significantly different at P < .05 (G-test) are represented
by different letters. Prey taxa are defined as follows: Hym-F, ants; Hym-NF, non-ant Hymenoptera; Dip-N,
nematocerous Diptera; Dip-NN, non-nematocerous Diptera; Col, Coleoptera; Hem, Hemiptera; Thy, Thysanop-
tera; Hom, Homoptera. 3, Proportion ofH. oculata involved in various activities at each census period. Values
for each census period are summed over the 10 study days. CP = copulating, CT = courting, F = feeding, O
= other (i.e., perched without prey or flying).
648 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
o
VOLUME 102, NUMBER 3 649
smaller non-nematocerous Diptera prey 40-,
than the average size available to it, and
significantly larger Coleoptera. H. oculata
chooses significantly larger formicid Hy- 30-
menoptera than the overall average size of
prey captured, and smaller non-formicid
Hymenoptera and Araneida. This apparent 20-
size preference has also been observed in
other robber flies (Hespenheide and Rubke
1977, Hespenheide 1978, Shelly 1984a) 10-
The proportion of flies involved in the
four activities (copulation, courtship, feed-
ing, or other) at different times of the day
is shown in Fig. 3. Copulation and court- 200 400 600 800
ship tend to peak in mid- to late afternoon
Netradiation(watts/m-^)
and the proportion offlies feeding generally
A
increases through the day. linear regres-
sion shows a strong inverse relationship be-
tween fly densities and net radiation (Fig. 30
4; P < .000' r- = .54) and a weaker in-
verse relationship between fly densities and
25
ambient aperature (P < .0001; r- = .34),
although li.ese factors are correlated (P <
= ^ 20
.0001; r^ .67).
Differences in robber fly behavior be-
15
tween the shady and sunny transects further
support the hypothesis that net radiation is
the primary influence. The cycle of average 10
daily robber fly density in the sunny and
shady transects is shown in Fig. 5. During
the study period, robber flies in the shady
transect exhibited no significant change in
average hourly density throughout the day 0800 1000 1200 1400 1600
whereas those in the sunny transect de- Time(hrs.)
creased significantly at midday comparedto
dawn or dusk. By 1400 h direct sunlight Figs. 4-5. 4, LinearrelationshipbetweenH. oculata
density and net radiation: as net radiation increasesfly
had entered the shady transect due to the density decreases (P < .0001; r- = .54). 5, The daily
sun's position at that hour, and the affect of cycle ofH. oculata in the sunny transect (opencircles)
this is represented in Fig. 5 as the period in and shady transect (closed circles). Data are shown as
the shady transect during which the fewest the mean density, ± SE, for each offive time periods
for 10 separate days. Different lettersrepresent signif-
flies were counted. Flies in the shady tran- icant differences among times within transects at P <
sect tended to congregate in shady patches .05 (Wilcoxon sign rank test); letters a-c refertocom-
when sunlight entered the area, or relocate parisons within the sunny transect, and letters d and e
out of the transect. In the sunny transect, to comparisons within the shady transect.
the few flies that chose to remain when net
radiation was highest were found perched flies may be responding primarily to inci-
under grass blades, positioned in such a dent radiation. Robber flies in general are
way so that the blade blocked the sun's believed to relocate to more shaded perches
rays. These observations indicate that the presumably to maintain a more "normal"
650 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
body temperature (Lavigne and Dennis may be an artifact of their relative abun-
1975). dance being higher in these margins, which
Fourteen pairs of H. oculata involved in appears to be the case. Alternatively, H.
courtship were observed in detail during oculata exhibits a species-specific perch be-
this study with five resulting in the male havior (see Fig. 1) like many asilids (Fisher
joining the female in copulo. Males locate and Hespenheide 1982, 1992), and grasses
females by slowly cruising through the hab- may present the optimal substrate for this.
itat close to the ground, pausing briefly to Ectatomma ruidum seized the robberflies
inspect perch sites for females. When a immediately in all 10 of the feeding trials.
perched female is located, the male hovers I conclude from this that, given the oppor-
several centimeters behind and slightly be- tunity under non-experimental conditions,
low the perched female and begins to fly the ant would prey upon H. oculata.
upwards and make physical contact with Of 62 prey items removed from 18 orb-
her. Each time the male makes contact, the weaver webs, 47 were ant reproductives,
female's wings are pushed apart; possibly nine were flies, and six were beetles. Inter-
to provide him with an exposed abdomen estingly, these orders also dominated the
to hold on to for leverage in initiating cop- prey sampled from Micrathena schreibersi
ulation. These collisions happen, on aver- (Perty) webs in Panama (see Shelly 1984b).
age, at about seven second intervals be- None of the orb-weaver's prey in my study
tween which the male resumes the hovering were robber flies. The absence of robber
position, rubbing his tarsae togetherpriorto flies as orb-weaver prey might be a result
the next attempt. Once in copulo the pair of robber flies cruising through the habitat
remains attached, tail to tail with the female at a slow enough rate as to evade becoming
perched and the male dangling upside ensnared. I observed one case in which a
down, for an average of 11 minutes (n = robber fly collided with a web and imme-
5). The female will often relocate to another diately flew around it. Non-nematocerous
perch, towing the male behind her. The nine Diptera are strong fliers and as agroup have
unsuccessful courtship attempts ended with been found to be disproportionately repre-
the female, apparently uninterested in sented in webs of M. schreibersi (Shelly
courting the male, flying far enough away 1984b), and this may likely be due to their
such that he lost track of her. The copula- ability to both avoid and escape becoming
tory position I describe here for H. oculata caught. This may explain the absence ofH.
is similar to that ofH.fusca in North Amer- oculata in orb-weaver webs during my
ica (Dennis 1979). study.
Among the 80 H. oculata observed dur-
Acknowledgments
ing the study of perch preference, most
chose grasses (90%) over woody branch I am grateful to Seth Bigalow, Ankila
tips, herbs, and sedges in and around the Hiremath, and John J. Ewel for allowing me
study area, with a mean ± SE perch height to work in the study site ofthe Huertos pro-
of 45.5 ±1.7 cm. The grass species and ject (NSF grant DEB-9318403) and forpro-
their order of perch preference included P. viding the data on net radiation. I am also
conjugatum (39%), Digitaria sp. (24%), C. grateful to the principal investigators and
dactylon (18%), and undetermined blades staff of the Arthiopods of La Selva project
(19%). Among the major parts of an indi- (John T Longino, Robert K. Colwell, Dan-
vidual grass, 60% of robber flies preferred ilo Brenes Madrigal, Ronald Vargas Castro,
the tips of spikes, 28% the main part of the Maylin Paniagua Guerrero, and Nelci
blade, 9% the tip of the blade, and 3% the Oconitrillo Miranda; NSF grant DEB-
stem of the spike. This clear preference for 9706976) for allowing me to use the facil-
grasses, and for P. conjugatum in particular, ities contained there. I thank Grant L. Gen-
VOLUME NUMBER
102, 3 651
try for first introducing me to H. oculata, Hespenheide, H. A. and M. A. Rubke. 1977. Prey,
and Henry A. Hespenheide and A. G. Scar- predatory behavior, and the daily cycle of Holo-
brough for reviewing the manuscript and pogon wilcoxi Martin (Diptera: Asilidae). Pan-Pa-
cific Entomologist 53(4): 277-285.
offering valuable comments. Eric Fisher Lavigne, R. J. and E R. Holland. 1969. Comparative
provided the final determination forH. ocu- behavior of eleven species of Wyoming robber
lata. Funding support during this project flies (Diptera: Asilidae). University of Wyoming
was supplied by the Vice-Chancellor for Agricultural Experimental Station Science Mono-
Research at UCLA, Kumar Patel, and the graph 18: 1-61.
Lavigne, R. J. and D. S. Dennis. 1975. Ethology of
Organization for Tropical Studies (OTS).
Ejferia frewingi with a description of the eggs
(Diptera: Asilidae). Annals of the Entomological
Literature Cited Society of America 68: 992-996.
McDade, L. A.. K. S. Bawa, H. A. Hespenheide and
Dennis, D. S. 1979. Ethology of Holcocephalafusca G. S. Hartshorn,eds. 1994. LaSelva: Ecologyand
in Virginia(Diptera: Asilidae). Proceedingsofthe Natural History ofa Neotropical Rain Forest. The
Entomological SocietyofWashington 81(3): 366- University ofChicago Press, Chicago, 486 pp.
378. Mensah, R. K. 1996. Evaluation of coloured sticky
Fisher, E. M. 1983. Pilicaformidolosa (moscaasesina, traps for monitoring populations of Austroasca
robber fly), pp. 755-758. In Janzen, D. H., ed., vividigrisea (Paoli) (Hemiptera: Cicadellidae) on
Costa Rican Natural History. The University of cotton farms. Australian Journal of Entomology
Chicago Press, Chicago, 816 pp. 35(4): 349-353.
Fisher, E. M. andH. A. Hespenheide. 1982. Taxonomy Scarbrough, A. G. 1982. Coexistence in two species
and ethology of a new Central American species of Holcocephala (Diptera: Asilidae) in a Mary-
ofrobber fly in the genus Glaphyropyga (Diptera: land habitat: Predatory behavior. Proceedings of
Asilidae). Proceedings of the Entomological So- the Entomological Society of Washington 84(2):
ciety ofWashington 84(4): 716-725. 349-365.
. 1992. Taxonomy and biology of Central Shelly, T. E. 1984a. Prey selection by Atractia mar-
American robber flies with an illustrated key to ginata. Proceedings ofthe Entomological Society
genera (Diptera: Asilidae), pp. 611-632. In Quin- ofWashington 86(1): 120-126.
tero,D. andA. Aiello,eds.. InsectsofPanamaand . 1984b. Prey selection by the neotropical spi-
Mesoamerica: Selected Studies. Oxford Universi- derMicrathenaschreibersiwith noteson web-site
ty Press, New York, New York, 692 pp. tenacity. Proceedings of the Entomological Soci-
Hespenheide, H. A. 1978. Prey, predatory and court- ety ofWashington 86(3): 493-502.
ship behavior of Nannocrytopogon neociilatus . 1988. Relative abundance of day-flying in-
Wilcox and Martin (Diptera: Asilidae). Journal of sects in treefall gaps vs. shaded understory in a
the Kansas Entomological Society 51(3): 449- neotropical forest. Biotropica 20(2): 114-119.
456. Weeks, L. and H. A. Hespenheide. 1985. Predatory
. 1989. Notes on the biology ofEjferia wilcoxi andmating behaviorofStichopogon (Diptera: As-
(Bromley) (Diptera: Asilidae). The Pan-Pacific ilidae) in Arizona. Pan-Pacific Entomologist
Entomologist 65(4): 375-380. 61(2): 95-104.