PHOTINUS COLLUSTRANS: REPRODUCTIVE ECOLOGY OF
FLIGHTLESS FEMALE FIREFLIES
STEVEN RAE WING
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Steven Rae Wing
Earth angel, earth angel
will you be mine?
I am grateful to my graduate committee members for their guidance,
patience, and friendship. The chairman, Dr. James E. Lloyd, introduced
me to fireflies and to the behavioral ecology context in which they are
so valuable. Throughout this project I depended on his extensive
knowledge and his gift for looking at problems from new perspectives.
Dr. Thomas J. Walker's work provided inspiration in developing these
studies, and his practical suggestions regarding methods facilitated
data gathering. Dr. Donald A. Dewsbury added another dimension to the
committee, in part because of his work with vertebrates, and his counsel
proved to be very beneficial. I am indebted to these committee members
for invaluable encouragement, constructive criticism, and for providing
inspiring examples of scientific integrity and achievement.
I owe much to my friends, Dr. John Sivinski, Susan A. Wineriter,
and Dr. Timothy G. Forrest. John and Susan never failed to infect me
with their enthusiasm, which helped to sustain my own. John tirelessly
reviewed first drafts of manuscripts and captained the plunder of the
bio-treasure buried in Fire Ant beds and other such adventures that kept
the fun in science. Tim took the photograph for Fig. 2 and gave freely
of his time and expertise to coerce a computer into producing Figs. 14,
16-22, 24-27, and collaborated with Susan to generate Fig. 13. Susan,
who was also generous with her time and talent, produced Figs. 12, 15,
I also benefited from contributions from other friends and
associates. Barbara Hollien worked with admirable energy and accuracy
in preparing the printed manuscript and revisions. Laura Line Reep drew
Fig. 29. Figure 29c was made from an unpublished drawing by Lloyd and
Ngo Dong. Warren Prince and Sara Lawrence both allowed me to cite
unpublished data. Dr. Will Hudson and Dr. Simon Chew each gave advice
on statistics. Identifications of specimens were provided by Will
hemipterann) and G.B. Edwards (spider). Dr. David Hall kindly visited
the site to identify vegetation. Dr. Scott Sakaluk made valuable
constructive comments on manuscripts. Dr. Richard Bruce facilitated the
study at Highlands Biological Station. Dr. Martin D. Young helped me
explore some different areas of science while providing employment that
financed part of this dissertation. I also acknowledge the Division of
Human Resources for the employment they provided, and my many friends
there for their encouragement and good humor. This work was also
partially funded by Dr. Lloyd's N.S.F. Grant #DEB-7821744, and by
Department of Entomology and Nematology funds (assistantships).
I am profoundly grateful to my family, Susan Schott-Wing, Jessica
Wing, and Sol Schott for their love, support, seemingly endless
sacrifice, and assistance. They made this project possible.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ............................................. iv
ABSTRACT ................................................. viii
INTRODUCTION ............................................. 1
GENERAL METHODS AND MATERIALS .............................. 7
FEMALE MATING FREQUENCY AND MALE COMPETITION
IN PHOTINUS COLLUSTRANS ................................ 9
Introduction .......................................... 9
Methods and Materials ................................... 9
Results ...................................... ...... 10
Discussion ...... ....... .............................. 12
COST OF MATING FOR FEMALE INSECTS: RISK OF PREDATION ......... 20
Introduction ........................................... 20
Methods ............................................... 21
Results and Discussion .......... ....................... 23
ENERGETIC COSTS OF MATING ................................ 29
Introduction ......................................... 29
Materials and Methods ................................ 31
Results and Discussion ................................. 32
TIMING OF REPRODUCTIVE ACTIVITY: FINDING A MATE IN TIME ...... 42
Introduction .......................................... 42
Methods and Materials .................................. 42
Results and Discussion .................................. 47
SUMMARY AND CONCLUSIONS ...................................... 81
EPILOGUE: COLLUSTRANS IN PERSPECTIVE ......................... 83
Introduction ............................................ 83
Materials and Methods ................................. 85
Results and Discussion .................................. 86
Conclusions ........................................... 92
LITERATURE CITED ............................................. 101
BIOGRAPHICAL SKETCH ........................................ 106
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PHOTINUS COLLUSTRANS: REPRODUCTIVE ECOLOGY OF
FLIGHTLESS FEMALE FIREFLIES
Steven Rae Wing
Chairman: James E. Lloyd
Major Department: Entomology and Nematology
Characteristics of Photinus collustrans fireflies, more fully
revealed here, allowed me to address topics in animal reproductive
ecology that had previously been investigated in few, or no, species in
Female collustrans are flightless and live in burrows in the soil.
They leave these burrows only for mating, and almost all females mate
only once. While outside of the burrow, females are more likely to be
attacked by predators than when in the burrow. The more nights it takes
for a female to mate, the fewer eggs she will deposit.
On the first night a female leaves her burrow, her exit coincides
with the nightly peak in numbers of searching males. On subsequent
nights, females that are still unmated exit earlier and so are available
during more of the male searching period. Females that are
unmated at the end of the male searching period for that evening may
remain outside the burrow long after male search has ended.
The brachypterous females of collustrans are less mobile and more
semelparous than either the alate females of other Photinus species or
the larviform females of Phausis reticulata. These chapters provide a
detailed description of an extremely limited adult role in a female
insect. This provides a base line case for comparison with other insect
there are two tribes of biologists. One is
devoted to solving some fundamental problem or other by
looking around for the right organism with which to
accomplish the trick. The other, to which I belong, is
devoted to glorifying their favorite organism by
finding important problems that it is ideally suited to
E. O. Wilson (1986)
For as long as there have been men, they have been watching
animals. With mankind's increasing understanding and manipulation of
the environment, the way in which observing animals is of value has
changed. Once men were concerned with avoiding predators, and with
catching prey. Today our survival depends no less on our knowledge of
nature, but most people are removed from active pursuit of greater
understanding of the natural world, and those who do pursue it are far
The study of insect behavior is one such specialization. The
activities of insects are of vital interest to mankind, for the effects
of insects range from pollination of crops to transmission of diseases
deadly to humans. Moreover, the topic is of interest in increasing our
understanding of our own animalness, and of the behavior of animals in
One aspect of animalness about which questions have been asked for
as long as there has been language is sex. Science, especially since
Darwin, has refined our concepts about sexual behavior. Darwin (1859)
established the concept of evolution by natural selection. His
successors, particularly in this century, clarified the mechanisms of
adaptation and developed the powerful model of selection acting at the
individual or gene level (see Dawkins 1976). New questions about mating
systems were posed as a result of this conceptual revolution. The very
definition of species became a question of who mates with whom.
Observations of behavior (and other traits), previously colored by less
powerful concepts, are now more productively considered in terms of how
characteristics contribute to reproductive success (see Lloyd 1980 and
The studies presented here are concerned with production of the
next generation by certain fireflies. At the time I began this project
some of the recently posed questions in reproductive ecology had not
been addressed by field biologists for lack of a suitable animal. I
selected the firefly species Photinus collustrans Le Conte (1878) for
study because it appeared likely to prove useful for investigations of
some of these topics, particularly the subject of female multiple
mating. Because many female insects obtain a lifetime supply of sperm
with one mating, subsequent matings may not be required. Yet many
female insects do mate repeatedly (Smith 1984 and refs.). My goal was
to monitor females throughout their adult lives and document their
complete sexual histories, in the field. This seemed an achievable goal
with collustrans females because of their unusual lifestyle.
It was J.E. Lloyd who first explored the behavioral ecology of
collustrans, exploited their accessibily, and guided his graduate
students toward further study of the species. Lloyd's (1966) extensive
survey of the fundamentals of behavior and ecology in the genus
illustrates the value of the accessible signal channel that most
fireflies, including collustrans, share. Bioluminescent signals can be
observed, recorded, and analyzed far more readily than chemical signals,
for example (as Lloyd is fond of pointing out), and some are useful
taxonomically (Lloyd 1966). But collustrans has additional attributes
that make the species unusually valuable for field studies.
Members of many other firefly species are active at heights or in
habitats that make them difficult, if not impossible, to observe at
close range. But collustrans is a grassland species, and males usually
fly between 0.5 and 1.5 m above the ground (Lloyd 1979b). Lloyd (1979b)
utilized this combination of low-flying males over a grassland habitat
by following nearly 200 individual collustrans males as they made
signaling flights in search of mates. Lloyd was able to quantify
distances travelled with a measuring wheel and record observations on
tape during these chases. The resulting portrait of "a day--or rather
evening--in the life" shows problems that males face and investments
they make in the search for mates.
Another illustration of collustrans accessibility is provided by
Adams' (1981, 1982) studies of male flight paths. Adams wanted to study
search paths in 2 dimensions. Lloyd suggested collustrans as one
subject of the study because males search at a relatively uniform
height, and because it is possible to follow individual males. Also,
the collustrans season is long, so much data can be collected in one
year. As Adams followed individual males he dropped a marker under each
flash. He graphed the paths described by these markers and ran computer
analyses. The result was another notable contribution to field biology,
this one showing the changes in male searching strategy as the evening
Lloyd and Adams had taken advantage of the accessibility of
searching males, but females remained relatively unstudied. Preliminary
observations by Lloyd (1966, 1979b) and, later, Wing (1982) showed that
females too have characteristics that make them accessible for certain
Females are flightless and flash responses to flying males while
stationed on blades of grass near the ground, or on the soil itself
(Lloyd 1966). The open nature of the grassland habitat and the
relatively exposed locations of responding females make them easy to
find. Flash dialogues by which males locate females make pair-formation
As I found in a previous study of female movements (Wing 1982),
females are not only easy to locate, but also easy to mark. They are
pale, soft bodied, conspicuous, and relatively slow-moving. I found
females by using standard lampyrological procedures, locating their
flash responses to males or to male-like signals that I made with a
A female's signals give away her location, but to actually see a
female I used a flashlight. I could then mark the pronotum or elytra
with white ink. After replacing her on the spot where I found her, the
female would soon resume her responses to flashes. (Similarly, mating
pairs can be observed with light without seeming to disturb them.)
The results of marking were not entirely conclusive because the
marking substance was not permanent in moist soil. However, the fact
that collustrans females, like many other fireflies, can be manipulated
in such a way and will then resume their response behavior is another
aspect that makes them good field biology subjects. But there is
another characteristic of the species that is unusual and allows studies
that could not be made on most other species.
As Lloyd (1966) noted, collustrans have a nightly mate-locating
period of unusually brief duration (on the order of 18 minutes nightly
for males). This means that it is relatively easy to monitor a female
during the entire nightly period when she could attract a mate. By
contrast, the nightly mate-locating period for some other firefly
females may last eight hours or more (Wing 1985). But the whole nightly
mate-seeking activity of a collustrans female may last half an hour or
less. Not only is what a female does during the mate-locating period
readily observable, but one is able to monitor her throughout the
period. This is a key characteristic that allowed me to compile
complete sexual histories of individual females (see mating frequency
Questions about what females do (i.e., how many times they mate)
are not fully addressed without investigating why they behave in this
way (see Thornhill and Alcock 1983). In considering this question, I
examined two of the costs of mating, risk of predation and energetic
costs. These costs also would be predicted to influence the timing of
mating efforts (T.J. Walker 1983a), a study of which is also presented
In the Epilogue I suggest the beginnings of a way to organize
comparisons of female firefly reproductive strategies. The
accessibility of fireflies in general, as noted above, and the diversity
within the family make such comparisons potentially among the most
instructive of any animal group. The studies presented here may prove
to have their greatest scientific value as a case for comparison. These
studies are also of value in that they address ideas about what may
occur in nature with evidence collected in the field. Such field
studies will contribute to formulating more accurate models of the
natural world, particularly regarding insect mating systems, and may
have unexpected value in other areas of biology (see Ballantyne 1987a,b
for an example of how understanding firefly mating was of major
significance in studies of the taxonomy of the group).
GENERAL METHODS AND MATERIALS
Methods and materials that were used in all of the subsequent
chapters are presented here.
Field studies of Photinus collustrans were conducted at a site
north of the Gainesville Regional Airport in Gainesville, Alachua
County, Florida (NW 1/4 Section 24, TWP R20E, T9S). The site was a
grassy field, mostly Axonopus affinis Chase (common carpetgrass), and
Eremochloa ophiuroides (Munro) Hack (centipede grass). Most of the
females were found on one section of the site under scattered trees
(mostly Quercus spp., Pinus spp., and Myrica cerifera). The remainder
of the site was treeless. Two streetlights sometimes illuminated parts
of the site. Females found in illuminated areas were not considered in
I located females by their flash and/or glow responses to penlight
simulations of male mating signals (see Lloyd 1966). The location of
each female was marked by a numbered flag placed about 15 cm to the
north of her. Flags were 10 x 40 cm strips of plastic held in place
with nails. Females use the same burrow throughout their adult lives2
1Lloyd (pers. comm.) has been observing fireflies at this site for over
a decade, and most of the collustrans females that he has found were in
the same restricted area where the females in this study were located.
2During this study 91 females were individually marked with Tech Pen Ink
dispensed from Hamilton's paint pots (T.J. Walker and Wineriter 1981).
Females appeared for up to 10 consecutive nights. Every appearance by
each marked female occurred at her original position.
(Wing 1982), and numbered flags were sufficient to identify each
individual. Copulations were timed with a stopwatch, and some were
observed with a magnifying glass.
To establish that individual animals contribute only one point to
the data set (see Machlis et al. 1985) generally requires that they be
identifiable as individuals. In these studies females were generally
identified as individuals, males were not. Specific comments are
included in the text.
FEMALE MATING FREQUENCY AND MALE COMPETITION
IN PHOTINUS COLLUSTRANS
Although the consequences of multiple mating by female insects have
been discussed at length (Smith 1984 and refs.), few studies have been
published on female mating frequency in the field. This is partly
because such activities are difficult to monitor under field conditions.
However, as noted above, matings by Photinus collustrans females can
easily be monitored in the field. The brachypterous females live in
burrows and remain near them. About 20 min after sunset males start to
fly and search for females, which take positions on the soil surface or
on vegetation (Lloyd 1966). Females flash in response to the signals of
flying males, which locate females by their responses. Each night
sexual activity is restricted to a period about 18 min long (Lloyd 1966,
also see T. Walker 1983a for a discussion of such 'sprees'). Female
collustrans live about 10 days after their first appearance (Wing 1982);
by observing a female for about 20 min per night for 10 nights, every
sexual activity of her life can be recorded. There is no evidence that
females mate under circumstances other than those mentioned above and
some evidence that they do not (next chapter).
Methods and Materials
An 18 x 20 m area was searched for females nightly starting before
male flights began and ending after they had ended for the night. I
marked the location of each female and I checked her position at
approximately 1 min intervals.
Each time a female's position was visited on a given night, she was
presented with penlight simulations of the male signal. Because females
do not respond following a successful mating, but instead pause and then
re-enter their burrow (Wing 1982), a female answer to my signal
indicated that she was not yet mated. If she failed to respond, I
determined whether she was (1) still at the burrow entrance but not
responsive, (2) mating, (3) entering her burrow, or (4) gone. All
female locations were checked until the adult season was over.
Because the same area was searched nightly throughout the season,
when a new female appeared she was almost certainly a virgin making her
first appearance. I recorded complete sexual histories of 108
Of the 108 females whose complete sexual histories were determined,
104 mated only once. The general sexual pattern was as follows. The
female appeared by her burrow nightly until she attracted a male (x = 2
nights). The male, having located the female by her continued responses
to his signals, landed nearby and walked to her. Upon making physical
contact, the male climbed upon the female and copulated with her in the
male-above position (Fig. 1). Copulation lasted about 1 min (details
below), and then the male broke the connection, dismounted, and flew
away leaving the female outside her burrow. Following copulation the
female paused for seconds or minutes, and did not flash responses to
signals of passing males. She then entered her burrow.
Four females mated more than once. One of these females was dug
from her burrow by a male, one was mated by a "sneaky" male, and two
made themselves available to other males by their own behavior.
The two repeated matings due to female behavior occurred when
females mated and entered their burrows, but on subsequent nights left
their burrows, responded to male signals, and mated again. Only three
of 108 females responded to signals on nights subsequent to the first
mating. Two mated again of which one remated once, the other twice.
When more than one male landed at a responding female, the first
male to reach her mounted and began copulation. The rival male
attempted to mount the female (sometimes backwards, see Lloyd 1979b),
and to break the pair apart (Fig. 2). As a result, the copulating male
moved or was pushed off the female and the copulation proceeded in the
tail-to-tail position, with variations. (In the tail-to-tail position
the male and female face in opposite directions while maintaining
genitalic connection (see Wojcik 1969). Due to disturbance by rival
males, pairs were sometimes moved into odd positions, even with the
female on the copulating male's dorsum.)
Copulations were significantly longer when rival males were
present. Mean duration for single male copulations was 57 sec (n = 23,
range 30-185 sec) compared to 842 sec (n = 5, range 339-1410 sec)
(Mann-Whitney (U = 115) P <.0005) (Zar 1974) when rivals were present
(also see Rutowski and Alcock 1980). In these cases, copulating males
maintained the genital connection until after the females had entered
1Males were not identified as individuals. However, intense competition
for females among numerous males makes it unlikely that one male mated
with more than one female (see Lloyd 1979b). Ideally, data would be
paired for individual males each mating with and without other males
their burrow (Fig. 3). Females entered head first, dragging the coupled
males backwards down the burrow. In one case, only the head and thorax
of the male remained outside the burrow when genital connection was
broken. After disengaging, males climbed out and flew away. After the
mated male departed, in four of seven cases the rival male dug at the
burrow opening (i.e., tried to remove the female from her burrow).
Rival males located the burrow opening by antennating the soil. Rivals
dug at the burrow (Fig. 4), sometimes completely entering it. On one
occasion the male succeeded in removing the female from her burrow and
mated with her (this accounted for the third multiple mating) (Fig. 5).
Unsuccessful males dug for as long as 35 min before leaving. In some
cases more than one rival male dug at the burrow (Fig. 6).
The fourth repeated mating resulted from another behavior of rival
males and was observed once during this study and once since then. The
rival "sneaky" male was non-aggressive, and made only occasional contact
with the copulating pair. The rival gently antennated the pair and then
walked away, returned and antennated the pair again. The copulating
male stayed in the male-above position, and copulation was not
prolonged. After copulation, the male dismounted and flew away, leaving
the female outside her burrow. The rival male then located the female
and mated with her.
A large investment of search time is likely to be required for a
male to find a responsive female (Lloyd 1979b). With a period of only
about 18 min nightly in which to operate, the usual male strategy after
locating a female is to mate and, within a minute or so, return to the
air searching for another female.
Females pause after mating but do not answer the flashes of
passing males. They then enter their burrow. Fewer than 3% of the
females that mated made themselves available to males again (also see
Wiklund 1982). Whether or not remating by collustrans females is
adaptive (see W.F. Walker 1980) is unclear.
During the pause before re-entering the burrow females are
susceptible to another mating if found by another male. Even after
re-entering the burrow a female may be dug up and remated. Generally,
then, if a male can gain physical access to a female, he can mate with
her. This fact has led to prolonged copulation when a rival male is
present (see Parker 1970). Copulating males make the female physically
unavailable by occupying her until she has returned to her burrow (see
Sivinski 1983). Rival males try to break the coupled pair apart and
attempt to gain access to the female by digging her from the burrow.
The "sneaky" rival avoids triggering mate-guarding by the coupled male.
and thereby gains access to the female after her first mate leaves.2
The complex of male strategies and counter-strategies shown here
reflects how important the potential for female multiple mating can be,
even when only a small proportion of females actually mate more than
once. If the number of observed matings was increased enough, it seems
likely that other strategies would be revealed, and exceptions to almost
any of the usual behavioral patterns would be found.
It should also be noted that mated females are sometimes flooded from
their burrows. and may remate under these circumstances (Wing 1982 and
Fig. 1. Photinus collustrans male (above) mounting female. Note
burrow opening in background. The female is about 11 nm long.
Fig. 2. Photinus collustrans trio. Male on left is copulating.
Rival has mounted female.
Fig. 3. Rival male (left) remains mounted as female enters burrow.
Male on right is copulating.
Fig. 4. Rival male digs at female's burrow.
k.e' ~v ,-~'
Fig. 5. Rival male pulls female from her burrow.
Fig. 6. Two males simultaneously dig at a burrow.
COST OF MATING FOR FEMALE INSECTS:
RISK OF PREDATION
The cost of mating includes those expenses and dangers
". involved in locating mates, in courtship, in competition with
rival suitors, and in copulation ." (Daly 1978, p. 771). One of the
costs discussed by Daly is increased risk of predation associated with
mating. Such predation is expected to influence the evolution of mating
behavior (W.F. Walker 1980, Greenfield 1981, Burk 1982, T.J. Walker
1983a), and the concept of this mating cost is supported by examples of
predators that orient to the mating signals of male insects (Cade 1975,
Lloyd and Wing 1983). However, for females only circumstantial evidence
of this mating cost exists (Sakaluk and Belwood 1984; also see Verrell
1986a). In fact, better evidence suggests that some female insects are
actually protected from predators during mating (Sivinski 1983; also see
Verrell 1985a concerning males). This chapter provides the first
empirical evidence, for any animal species, that females experience an
increased risk of predation as a consequence of mating (see Sakaluk and
Cade 1983; Thornhill and Alcock 1983).
Assessment of this cost of mating for females has been hindered by
the difficulty of determining precisely when females are engaged in
mating activities. However, the reproductive behavior of a Photinus
collustrans firefly female can be unambiguously distinguished from
other biological activities. These females are flightless and inhabit
burrows in the soil. A burrow is occupied by a solitary female, and she
does not stray from the immediate vicinity of the burrow's opening
(females do not move from one burrow to another, preceding chapter).
During the brief period shortly after sunset when males signal, females
take positions just outside of their burrows and flash responses to the
signals of flying males. By flash dialogues males locate females and
copulate with them (Lloyd 1966, 1979b). Daly (1978) suggested that such
signals must disclose the locations of females to predators as well.
Mating costs are revealed by contrasting ". the costs incurred
by a sexually reproducing female with those incurred by a hypothetical
asexual reproducer" (Daly 1978, p. 771). Females of P. collustrans must
leave the burrow to mate, thereby exposing themselves to predators found
above ground. However, a hypothetical asexual female would not have to
leave the burrow at all (below). The risk of predation associated with
mating can be demonstrated, then, by showing that the risk of predation
is greater outside of the burrow than in it.
To determine whether females leave their burrows at times other
than during the male signaling period, I used two balsawood splinters
(each with a black ring inked around its center). These were aligned to
form crosshairs over the burrow opening. The crosshairs were disturbed
in a telltale way by a female leaving (or entering) her burrow.
Crosshairs were placed over the burrow after a female (of known mating
history, methods explained in preceding chapter) retired for the
evening. Data from 15 different females showed that, prior to mating,
females left the burrow only pursuant to mating, and mated females did
not leave at all. The same was true of 14 females housed in vials with
soil. These females appeared more or less nightly until mated, and
thereafter stayed underground where they oviposited and died (Wing
1982). (A small percentage of mated females do return on subsequent
nights for a second mating, preceding chapter.) Thus, crosshairs make
it possible to detect the disappearance of a female, and to ascertain
whether she was in or out of the burrow when she disappeared. Because
females never leave the immediate vicinity of the burrow under normal
circumstances (preceding chapter; Wing 1982), disappearances can be
safely attributed to predation.
Females spend most of the day in their burrows. Males fly during
a period of about 18 minutes duration nightly (Lloyd 1966), and females
usually stay out less than an hour (see chapter on timing). However,
for convenience, the time spent outside the burrow by a female on one
night was rounded up to one hour, and the daily time spent in the burrow
was counted as 23 hours. (These figures err in favor of the null
hypothesis, i.e., risk does not increase for the female outside of the
Data were recorded as follows. When a female appeared, crosshairs
were placed over the entry of her burrow. If the crosshairs were moved,
indicating that the female re-entered, the hour was recorded as "safe."
However, if she disappeared while out of the burrow crosshairss not
moved, indicating no re-entry) a lethal attack was presumed and the hour
was recorded as "terminal." Also, when a known predator of these
1Females may be forced from the burrow by flooding.
fireflies, a lycosid spider (Fig. 7, see Lloyd 1973), was seen attacking
a female but apparently was frightened away by my light, a terminal hour
was counted. (Predators suspected in the disappearances of other
females included ants, a toad (see Lloyd 1979b), and reduviids.2)
Mortality in the burrow was implied when a female known to be
unmated failed to leave her crosshaired burrow on subsequent nights.3
In such a case, 1 terminal hour was recorded. If instead the female did
emerge the next evening, 23 safe hours were counted.
Results and Discussion
The results are shown in Table 1. Of the 944 hours monitored in
the burrow, one hour was terminal. Outside of the burrow, 6 of 53 hours
were terminal. The proportion of terminal hours in the burrow (1/944)
was used to generate the number of terminal hours expected outside the
burrow (1/944 x 53) if the risk of mortality does not increase. A
2The reduviid Repipta taurus frequently preys on collustrans males. In
captivity one readily attacked a collustrans female. Photuris females
also prey on collustrans males and are found on the ground at the site.
3One immature spider (Lycosa lenta) was found exiting from a female's
burrow, but whether the tiny spider could prey on the much larger female
4Females were monitored where they naturally occurred, scattered in
space and time. Female collustrans are unlikely to vary much in ability
to detect and avoid predators because females seem uniformly oblivious
to other creatures (except flashing males) until physically disturbed,
and they appear to be equally defenseless. The females I monitored were
of similar ages and sizes, so presumably similarly attractive to
predators. For these reasons I believe that the proportion of terminal
hours in the burrow reflects the distribution and abilities of
predators, giving a reasonable estimate of how often females in their
burrows are found and attacked. However, the method of collecting data
is not without risk. Females contributed more than one observation
(i.e., hour) to the data set (23 hours for 1 safe night). This may
constitute Pooling, and if so, statistical inferences based on the data
could be jeopardized (Machlis et al. 1985).
Table 1. Comparison of safe vs. terminal hours spent by collustrans
females in vs. out of the burrow.
in out out Z P
Terminal 1 6 0.06 2.36 < .01
Safe 943 47 51.94 2.36 < .01
similar method was used to generate the expected number of safe hours
outside the burrow. The one-tailed binomial test (Zar 1974) was used to
compare the expected with the observed terminal and safe hours outside
the burrow. For these females, the increase in mortality risk
associated with being outside of the burrow for mating is statistically
highly significant (P <.01).
Such increased risk is expected to influence the evolution of
female mating behavior. When mating involves increased danger, females
may minimize the time spent in courtship and/or copulation. In
collustrans, this means minimizing the time spent outside of the burrow.
Three aspects of their behavior may be involved. First, exits of
females from their burrows coincide with the peak availability of males
(see chapter on timing). Thus, females are outside when they have the
best chance of being seen by a potential mate. Second, copulation
durations of collustrans pairs are remarkably brief by known firefly
standards (see Epilogue). Other fireflies, with more vagile females,
may couple for hours or days, but usually collustrans matings last about
1 minute (preceding chapter; Wing 1985). Third, although females of
other firefly species are known to mate repeatedly5 (Wing 1985; Sara
Lewis, unpublished), collustrans females generally copulate only once
(preceding chapter). As a result of these 3 characteristics, a
collustrans female may leave the burrow only once in her adult life and
return to the burrow within a few minutes carrying her store of sperm.
More extreme still, subterranean females of at least one other firefly
5Females have never been observed to plug the entrance to the burrow
even after mating. This is compatible with the notion of relative
safety in the burrow, although it does not necessarily support the idea
because the burrow may be left unplugged for other reasons.
species, Lucidota luteicollis, may completely avoid coming above
ground for mating. These females sometimes mate through the sand,
apparently attracting their aboveground males with pheromones instead
of light emissions (Warren Prince, unpublished; Lloyd, unpublished).
This signaling channel may have evolved in lampyrids in response to
predation (see Greenfield 1981). Of course, many other factors may
also influence the evolution of reproductive characteristics such as
these (Thornhill and Alcock 1983).
ENERGETIC COSTS OF MATING
All cold blooded animals spend an
unexpectedly large proportion of their
time doing nothing at all, or at any
rate, nothing in particular.
Charles Elton (1936)
Daly (1978) noted that there are energetic costs of mating.
"Besides the costs of producing structures (including behavioral
substrates), the actual performance of behavior costs time and energy
too" (also see Verrell 1985b). Female collustrans invest some of their
resources in mating-related structures, including compound eyes and a
sperm storage organ, and invest time and energy in excursions above
ground in the quest for a supply of sperm. Hypothetical asexual
fireflies could avoid these energetic costs of mating, remain in the
soil, and deposit their eggs without delay. The focus of this chapter
is the cost of delaying oviposition to accommodate acquisition of sperm.
The acquisition of nutritional substances by an animal is an
important component of energetic costs of mating. For example, some
female insects acquire substantial nutritional resources as a result of
mating. Some scorpionfly males capture prey which are presented to
females during courtship, and on which females feed while mating
(Thornhill 1980). In many Orthoptera, spermatophores of considerable
food value are passed to females, which eat them (Gwynne 1980). In
other insects, the spermatophore is apparently digested within the
reproductive tract (Boggs and Gilbert 1979). In these cases, sexual
females could produce more eggs than hypothetical asexual females that
did not enjoy the nutritional benefits of mating. However, collustrans
females seem to be at the other extreme. They are among the least
likely fireflies to gain nutritionally from mating, judging from
comparisons of firefly male accessory glands and mating durations (Wing
1982, 1985), lack of female multiple mating, and the fact that no
nuptial feeding has been observed (see Epilogue).
Furthermore, I have tentatively concluded that collustrans females
do very little, if any, feeding as adults. In observations of hundreds
of females out of their burrow for mating, I have seen no feeding, and
there is evidence that females do not feed while in the burrow either.
Larvae that closed from eggs deposited in "ant farm" burrows by mated
females survived for as long as a year by actively hunting and feeding
on small earthworms (Fig. 8) in the soil. However, females remained
stationary while in the burrow, and did not hunt prey underground.
Unmated females remained facing the burrow entrance, rarely moving
except to leave in the evening and to return. Mated females moved
around in the burrow depositing eggs, and shortly thereafter died
Female collustrans, then, apparently enter adulthood with certain
metabolic resources which they do not replenish. A delay in oviposition
to accommodate mating must be financed from these resources, which
should result in the production of fewer (and/or perhaps smaller or
lighter) eggs (Yuma 1984, Forrest 1986, Karlsson 1987). This study
considers effect of delayed mating on egg count.
Materials and Methods
Mated and unmated collustrans females were collected in the field
and housed in the laboratory. Some were kept in cylindrical plastic
vials (3 cm diam x 2h cm deep, with lids) with soil from the site.
Others were housed in plexiglass-sided sandwiches similar to "ant
farms." These were constructed from 2 plexiglass panes spaced 6.4 mm
(W") apart with a U-shaped plywood frame. Through the top opening the
"farm" was filled about 3/4 full of soil from the site. A 3 cm deep
burrow-like impression in the soil was made so that one wall of the
burrow was plexiglass. On each side of the burrow a deeper impression
was made for adding distilled water with an eyedropper. A removable
plywood crossbar blocked the top. Cardboard covers were clipped in
place over the outside of the plexiglass except during daily
observations, some of which were made with a dissecting microscope.
The weight of some females was followed over time by weighing
them periodically on a Mettler digital balance accurate to .001 g. For
weighing, live females were cleaned under a dissecting microscope by
gently brushing off soil grains.
Some females of known weight were dissected and counts of eggs were
made. A computerized statistical analysis program, SAS Proc GLM, was
employed to analyze the relationship between female weight and number of
Other females were left undisturbed in the field and observed
nightly. The locations of these females were marked with numbered
stakes, so that individuals could be identified and monitored over time.
Females that had not succeeded in mating for several nights were
collected and dissected for egg counts, to compare with counts from
females collected on the first night they appeared.
Results and Discussion
On their first night (Fig. 10), females contained x = 71 eggs, n =
15 females, range 30 to 112 eggs, SD = 23. But females that continued
mating efforts, without success, for over a week were almost devoid of
eggs and grew very thin (Fig. 11). Two females were dissected, one
after attempting to mate on 7 consecutive nights, the other 10 nights.
They contained 10 and 9 eggs, respectively. Based on these data,
females seem to lose an average of about 7 eggs per day.
In the laboratory, egg "loss" was followed by weighing females
periodically. The validity of this approach was shown by correlation
and regression on number eggs vs. weight. Based on the "wet" weights of
15 females of various ages that were subsequently dissected for counts
of full-sized eggs, there was a correlation coefficient of .93 for egg
count vs. weight. A regression on these data shows that the number of
eggs = .78 x weight in mg (F = 76.41, Pr >F .0001, R2 = .86 (Fig. 12).
Four females weighed periodically lost weight (and therefore eggs)
consistently over time (Fig. 13). Again, an average loss of several
eggs per day is indicated, although the rate of loss may change. These
unmated females did not deposit eggs in their vials, suggesting that the
eggs were metabolized to fuel continued mating efforts.
Entering adulthood with non-replenishable metabolic resources has
implications for how the reserves are used in the ultimate mission of
1Such females glow less brightly than younger females, which probably
makes them less conspicuous to males. (See Schwalb 1961 for a case of
aging females that glow more brightly.)
the female, egg production (see Fritz et al. 1982). Female collustrans
are among the more semelparous of fireflies (see Epilogue). When they
first appear, most collustrans females already have nearly all their
oocytes fully developed. Mated collustrans females housed in "ant
farms" deposited their eggs in the walls of the burrow over a period of
few days, and then died in the burrow.
The strong correlation of eggs and weight is another indication of
the semelparous extreme exhibited by collustrans females. Females that
ripen several clutches of eggs over time, such as some Photuris species
apparently do, would not be predicted to exhibit such a strong
correlation, especially if they periodically acquire nutritional
substances (e.g., prey). In this case a female's weight and egg count
could fluctuate through adulthood. A heavy female that had just
acquired a nutritional supply, but had not yet matured a clutch of eggs,
for example, could contain the same number of mature eggs as a female
with an empty stomach that had just oviposited and weighed much less.
But first night collustrans females carry their entire and only clutch
of eggs, and the longer the delay before oviposition the fewer eggs are
left to deposit.
Unlike the risk of predation, which is a possibility for these
females that are exposed during mating efforts, the loss of eggs in
prolonged efforts to mate is an inevitability, a certainty. Staying in
the burrow except for mating not only affords protection from predators,
but also conserves energy. The possible effects of the costs of mating
on the timing of female mating efforts will be discussed in the next
Fig. 8. First instar Photinus collustrans larvae feeding on an
.. ... .
&**''' '' 49 ^^, ^f ~
t-.^ -' '.-' -.I
*-/,. ', :,- -
. ,: -,j *. : U ., .:"
Y,. ,. ;. .-,
* "., ., ,- l '' ,9,
',.. .B. : ..,. l
.-- ;L ,"; .,' "" "
:.K _..,:"* ., .. '" i .C
r^-w,/v ;; ^" ". '< *S W 1' '
"-. R ^.^ ^
LvS^ *r' 'w S '. -'..* ^
Fig. 9. Photinus collustrans female in her burrow in the
laboratory. Note eggs in the soil below female.
Fig. 10. Typical Photinus collustrans female on the night she
Fig. 11. Photinus collustrans female on the 9th consecutive night.
On the night she first appeared this female looked like the female in
-- .-". ". .. ." .
^ I. -* J-^ -** "" **--
- o ^V^', N^
,\. ,' .
/. ... ". jl .o..t s:' ^-' t-- '- 1- -_
~ ~ : : -, 7~I ;,, ,- .,, ,.,.... ,. .
**" r ^ <- *- .: + ^+ ".-"- .:.i-'. *: r-- .. ^^ ,... : ^..
...+.'*-^ ^ ^ .. ...... : ra '' "' _* ^h.^j S_ ^'" '^..^" .. ...; "<
.- ""^ ^ ,.. ^ **" ^ ;,' *.-- ,- *" '^ .
. -3 .' ,
'f *; -._." t *. ; "- -.- \ *''.. '
,_ .,**' r-.'^
.... ,.. ,-.,
+..t~ L ,Q-'+ | r : it: ._+. ..., _,...
.. .....~~~~~~~~~L .,.S.,. .. .. ,
...,. ,. .., ,....# .:. .
~~~ .= rZ... '.+.,-:,.. ,- .<
.+.,+.-,. ,+,+ ", '+/ ;:. ,.-
P:.. ~, 2..".':<- ',-
.. ,< "--" ..." -+ ,-. ., .
+;i~ f~'~ L '".
- ..- ,.j .... ."
+; +++ ~ul >+,,: + .. ., ."C L'. ., ', .,-
..~~ ~ .. -;, .
Fig. 12. The relationship between the weight of a female and the
number of mature eggs she carries.
80 100 120
Weight of Female (mg)
0 Female A
0 Female B
A Female C
O Female D
Fig. 13. Changes in the weights of four females over time.
I I I I I I I I I I I I
1 2 3 4 5 6 6 7 8 9 10 11
TIMING OF REPRODUCTIVE ACTIVITY:
FINDING A MATE IN TIME
Absolute failure to reproduce can be avoided, in most sexual
insects, only if a male and a female find each other in time and space
and mate. This chapter is concerned with the temporal locating of
mates by Photinus collustrans adults. Two components of finding a mate
in time will be discussed. First males and females must synchronize
their adulthood, so that reproductive maturity occurs simultaneously in
the two sexes (seasonal synchrony). Second, males and females must be
sexually active at the same time of day (diel synchrony).
Methods and Materials
Field studies of collustrans were conducted during the years 1982
Maturity periods were determined during nightly visits to the site
from April to October during the years 1982 through 1984. During 1985
weekly visits were made. The presence of collustrans adults was
detected by their flashes. The start of male activity was recorded as
In some apterygote species, sperm transfer can be accomplished without
the meeting of male and female. Males deposit spermatophores on
substrates where females later encounter these sperm packets and become
inseminated in the absence of males (Schaller 1971; also see Turk 1988).
the time when the first male of the evening was seen flying and flashing
under the trees where females were most numerous. The last such flashes
ended the male activity period. The duration of the activity period was
thus the time between these first and last male signals.
An index of male abundance was obtained by counting the number of
signaling males crossing a transect during the activity period. This
was done approximately once a week during 1983 and 1984.
In 1983, the nightly timing of male abundance was determined as
follows. The time of the first male signaling flight under the trees
was noted. At the beginning of the next minute, I started a stopwatch
and began counting the flashing males crossing a 15.2 m (50 ft) transect
between two of the trees. The same transect was used all year.2 The
males that crossed the transect were counted on a hand-held mechanical
counter. The running count and the number of minutes that elapsed since
the appearance of the first male were recorded at the end of each
minute. The transect was watched for at least two minutes after the
last male flew across the transect, by which time there were no
signaling males flying in the area (under the trees). The number of
males crossing the transect during each minute after that evening's
sunset was calculated. These data were summed over the season to
produce a graph of the timing of male abundance in terms of number of
males crossing per minute after sunset.
The study of the timing of male abundance was repeated in 1984 with
the sample unit being instead the number of males crossing per 1/25 crep
(Nielsen 1961) rather than minutes after sunset. Crep is defined by
2Male activity occurs in the same area throughout the season.
Nielsen as the duration of civil twilight3 in minutes. The crep value
is 0 at sunset, and 1 at the end of twilight. Crep values were used to
correct for differing rates of the onset of darkness over the season,
and to make the results of this study comparable to similar studies at
other altitudes and/or latitudes, and/or seasons. In the 1984 study the
same methods and materials were used as in 1983 except that the
stopwatch was replaced with a programmable timer. The timer was
programmed to beep at intervals of 1/25 of whatever the crep unit for
that evening was. (The crep unit varied from 23 to 28 minutes over the
season.) The running count of the signaling males crossing a 6.1 m (20
ft) transect (the northern 2/5 of the 1983 transect) was recorded at
intervals of 1/25 crep. The data for the 1984 season were summed over
the season to produce a frequency distribution of male activity in terms
of crep units.
Although the 1983 male data collected in minutes after sunset,
discussed above, cannot be converted to crep units as accurately as data
on females (below), an approximation of the frequency distribution in
creps was produced as follows. The average of the duration of twilight
in minutes for the evenings on which males were counted was calculated
(= 26 min). The summed data of number of males crossing per minute
after sunset in 1983 were converted to number crossing per 1/25 crep
(= 1.04 min) by multiplying the fraction of the minute after sunset
overlapping a particular 1/25 crep by the summed number of males counted
in that minute. Adding together these numbers of males corresponding to
each 1/25 crep gave a total number of males crossing per 1/25 crep.
Civil twilight is the time between sunset and when the center of the
sun is 6* below the horizon.
The 1983 male data converted to approximate the temporal
distribution in crep units (above), and the 1984 data which were
recorded in crep units were combined to produce a generalized curve for
males as follows. To give each year equal weight, although 7.26 times
as many males were counted in 1983 as in 1984, each 1983 count was
multiplied by 1/7.26. The resulting 1983 counts were then added to the
1984 counts. These combined counts were used to calculate the more
meaningful percent of total males crossing.
Data on female seasonal (annual) abundance and the timing of their
availability were collected during the 1982 and 1983 seasons.
An 18 x 40 m area was searched for females nightly starting before
male flights began and ending after they had stopped for the night. The
location of each female was marked and I inspected her position at
approximately 1-min intervals (these inspections are elaborated on in
the Methods section of the chapter on mating frequency). Complete
histories of the sexual activities of individual females were
accumulated. This provided data concerning seasonal abundance and the
daily timing of the sexual availability of females of known 'ages.' On
the night of her first appearance a female's age was 1 night. If she
appeared the next night, her age was 2 nights, and so on. (Note that I
have not determined when, in relation to eclosion as adults, females
make their first appearances to answer males.) The nightly timing of
availability was summed over the season in temporal units of minutes
after sunset for comparison with the 1983 data on number of males
available in minutes after sunset. The data on female availability were
also converted to crep units for comparison with the combined 1983 and
1984 data on males available per 1/25 crep.
Because males were scarce in 1982, the timing of female
availability was studied again in 1983, a year when males were abundant.
Because the search for females could not be made every night (as it had
been in 1982) due to the weekly transect counts of males, the female
data were treated as follows. For data concerned with female age, a
female was assumed to be making its first appearance as an adult (age =
1 night) only if she was a female that had not appeared the previous
night and that looked young (the abdomen appeared plump and white,
Fig. 10). Females that were first found on the night following a count
of males were considered new, but age (and mating history) unknown, and
these were excluded from analysis that required age values.
The relationship between the number of males counted crossing a
20-ft transect on a given night to the number of males that a female
might see was investigated (also see Otte and Smiley 1977). In order to
determine whether a female positioned on a transect could see the flash
of a male flying ten feet down the transect, the following was done. A
collustrans female in the field was presented with flashes from a very
dim penlight (these appeared to my eye dimmer than male flashes). While
making these flashes I stood ten feet away from the female (as measured
on the ground) and held the penlight about 1A m above the ground. This
was repeated on 3 other sides ( n90 apart) of the female.
The response of females to the flashes of males of different
species was tested by presenting females with flash patterns of various
species made with a dim penlight. These tests were done when there were
no collustrans males around for the females to respond to.
Results and Discussion
Adult collustrans were found during the warmer half of the year.
The first adults were found in mid-April, and the last in late September
or early October (Fig. 14a-e; also see Lloyd 1966 and Adams 1982).
All other things being equal, the more males there are searching
for females, the fewer females will be found in the nightly sample.
This is because females are located by their flash responses, and they
cease to respond once mounted by a male (Wing 1982). This fact
complicates comparisons of absolute numbers of females found during
times of different levels of male abundance. Nevertheless, three
patterns in annual (seasonal) abundance are clear.
Bimodal population peak
First, there are two annual abundance peaks (as Lloyd's unpublished
census data recorded over the last two decades also indicate, pers.
comm.). This pattern is evident for both males (Fig. 14b,e; 1983) and
females (1982 and 19831 Fig. 14a,b,e). There is a peak in late April
and early May, and another from August to early September. During the
time between these peaks the population of adults sometimes drops to
zero (1981, not shown). However, the 1984 and 1985 data (Fig. 14c,d)
show that there are not always two annual peaks.
The second pattern evident from the data in Fig. 14 is large annual
fluctuations in population size. There was a severe drop in abundance
beginning in the summer of 1984 (Fig. 14c,d) and lasting at least
through 1987. Males were virtually absent at the site for most of 1985.
The total counted over the season was less than the number of males that
crossed a 20-ft transect on some individual nights in 1984. Similarly,
in 1982 over 300 females were found by searching one 18 x 20 m area
nightly. During 1985 (not shown in Fig. 14) only 21 females were found
during 27 weekly extensive searches of the entire locale, with almost no
competition from searching males.
The rainfall pattern probably contributed to this decline (also,
see Yuma and Ono 1985). Because collustrans females and immatures
inhabit the soil, they seem to be sensitive to both flooding and to
drought. Flooding left standing water on the female search area for
weeks during the summer of 1984, and no summer abundance peak occurred.
During such floods bioluminescent, soil-dwelling creatures including
collustrans females, Photinus spp. larvae, and phengodids (Wing 1984)
were found dead on the surface of the water or clinging to vegetation.
Following the 1984 summer flood, a drought occurred. Larvae are
presumably developing between October and March in North Florida
(voltinism is unknown). During this study u 20 inches of rain fell
before the 1983 adult season (recorded about 1.1 km from the site by the
Federal Aviation Authority), and 21 inches fell before the 1984
season. However, only u 8 inches of rain fell before the 1985 season,
and very few adults appeared. A similar drought preceded the 1981
season, when adults were also scarce (not shown).
Corresponding male and female peaks
The third pattern evident from Fig. 14 is that in both 1983 and
1984 (Fig. 14b,c) peaks of male and female abundance are more or less
synchronized. Males and females are most numerous during the same weeks
of the year.
Finding a mate in seasonal time is a developmental problem. An
insect must reach adulthood with members of the opposite sex. Immature
members of both sexes share the same habitat through the larval and
pupal stages. Thus, each has the same information (cues) available with
which to time its eclosion. As discussed below, once they become
adults, males and females no longer have identical access to timing
One final point regarding Fig. 14: the 1984 male data (Fig. 14c)
give a conservative estimation of the number of males a female would see
in a night if she were out for the entire male activity period (i.e., a
range of 0 to over 200 males per night). This was shown by testing
whether females could see (would respond to) male-like flashes 10 ft
down a transect. Of 5 females tested, all consistently responded to
signals produced on at least 2 sides.4 This indicates that a female
could see all the males crossing a 20-ft transect (as was used in 1984)
if she were positioned in the center. (See Lloyd 1979b for similar data
on males; also see Otte and Smiley 1977.) Changes in male density, such
as those shown for the 1984 season, can influence the time required for
a female to get mated (see Dreisig 1971 and Lloyd 1979b).
4As noted by other researchers (Forrest, pers. comm.; Lloyd, pers.
comm.; Sivinski, pers. comm.) females can often see a flash much farther
away than 10 ft. One of the females tested responded to flashes I made
while standing 25 ft away. Definitive research on the female field of
vision was beyond the scope of this study. A female should be observed
during testing to determine whether she changes position, but such
observation should not interfere with her ability to see the male-like
signals. I was unable to do this with my rudimentary equipment. In the
tests reported here, which were done in the field, obstacles such as
grass could have prevented females from seeing the signals at some
Timing of Nightly Sexual Activity
In collustrans, and perhaps most other species of luminescent
fireflies, it is the males that actively seek mates by making signaling
flights (Lloyd 1966, 1971; also see Greenfield 1981 and Burk 1982).
Female collustrans make themselves available by exiting from their
burrows, perching nearby, and responding to male signals (Lloyd 1966,
1979b; Wing 1982). What is unusual about the temporal ecology of
collustrans is the brevity of the nightly period during which
mate-locating occurs. Males of some other firefly species, for example
Photinus macdermotti, make signaling flights throughout the night (Wing
1985). The duration of the collustrans male activity period is one of
the briefest known, being limited to about 18 minutes nightly (Lloyd
1966). For collustrans females, it is critical to be available during
this same 1% of the (24 hour) day that males are active. This section
is concerned with how females achieve the temporal overlap with active
males. First, some details of the timing of male activity are examined.
The timing of female availability is then explored, and compared with
the timing of male mate-locating activity.
Timing of male activity
In this study the male activity period is defined as the time
between the first and last male signaling flight of the evening. The
period occurs soon after sunset, a time when the ambient light level is
rapidly changing (see Dreisig 1971). That ambient light is the cue that
male fireflies use to determine when to make signaling flights has long
been recognized (see Lloyd 1966; Dreisig 1974). Males in darker
(forested) habitats begin signaling flights before males in lighter
(open) habitats do (Lloyd, pers. comm.; Sivinski, pers. comm.; also, see
Dreisig 1974). In this study the males under the trees always started
before males in the adjacent pasture (sign test, n = 24 nights,
P <.0002).5 The nonparametric sign test utilizes only the direction,
or sign, of the difference between members of a paired sample (Zar
1974). In this case the paired sample is the start time of males over
the pasture and the start time under the trees on a given night. If
males start first over the pasture, a + is arbitrarily assigned, and if
they start first under the trees a -. The sign test is used to
determine the probability of the observed distribution of + and if the
null hypothesis is true. The null hypothesis is that males start over
the pasture at the same time as those under the trees (sign = 0), or the
frequency of positive differences is equal to that of negative
differences. Males under the trees also stopped before males over the
adjacent pasture (sign test, n = 15 nights, P < .001). Another
indication of male reliance on ambient light cues is their earlier
activity when the western sky is darkened by clouds (see below).
In any one habitat, e.g., under the trees, male activity might be
expected to start and stop at about the same time each night, and the
duration of activity should be constant from night to night. However,
there is variation in the timing of male activity. In 1983 durations
ranged from 9 minutes to 26 minutes (n = 57 nights, mean = 16 min, SD =
3 min) (also see Lloyd 1966). Such variability can be due to variation
in starting time, stopping time, or both. Starting time is considered
An individual male could have contributed more than one point to the
data set, either by being the first male in both areas within a night,
or by being a first male on more than one night. However, the
probability of either is slight owing to the large number of males.
In 1983 male start times (n = 72 nights) ranged from 9 to 22 MAS
(Minutes After Sunset) with a mean male start time of 15 MAS (SD = 3.5
MAS). Variation in male start time is partly accounted for by external
factors that influence the time at which the critical ambient light
level occurs. Two of these factors are cloud cover and time of year
(how fast the light fades after sunset) (Table 2).
However, the major source of variation in 1982 male start times was
apparently not abiotic, but biotic in nature. Start times within the
season occurred earlier when the density of searching males was high
(Table 2) (also see Lloyd 1966).
The cause of this correlation remains unclear. There is probably
some variation among individual males in what level of light triggers
signaling flight behavior (see Dreisig 1971). In a larger sample one
might expect to find more "outliers" and an earlier first male. An
alternative explanation is that males are observant of and influenced by
the behavior of other males (see Lloyd 1979, 1982). Adams (1981)
quantified differences in collustrans males' flight patterns depending
on male density. Thus, there is evidence that collustrans males do
observe other flashing males and modify their behavior accordingly.
The variation in male stop times was also analyzed by regression
(Table 3). The factors that make darkness come earlier, shorter crep
and increasing cloud cover, resulted in males stopping earlier (although
the significance level of clouds is marginal at n = 20). However, the
effect of male density transectt counts) on stop time was not
significant (Table 3).
In considering variation in duration, the effect of clouds cancels
itself out. When darkness falls earlier, males start earlier (Table 2),
Table 2. Sources of variation in male start times.
Variable F value Prob.> F
number 20.72 .0003
crep 5.73 .03
Rsquare = 0.71
clouds 4.63 .05
number -0.01 (Start,1 min earlier per additional 100 males.)
crep 0.95 (Start 57 sec later per min added to crep.)
clouds -3.62 (Start almost 4 min earlier if clouds 100% as
compared to 0%.)
Regression on dependent variable start time. Significance level for
entry into the model = 0.15. Model: start time = number crep *
clouds. Number observations = 20 nights. The variable "number" total
number of males crossing the transect that night. "Clouds" = estimated
% of western sky occluded. (Proc. Stepwise, SAS 1982.) "Crep" is
number of minutes in 1 crep.
Table 3. Sources of variation in collustrans male stop time.
Variable F value Prob. >F
crep 3.77 0.044
Rsquare = 0.31
clouds 2.91 0.106
Number did not meet 0.15 significance level for entry into the model.
crep 0.88 (Stop '53 sec later per min added to crep.)
clouds -3.28 (Stop over 3 min earlier if clouds 100% as
compared to 0%.)
Regression on stop time. Significance level for entry into the model =
0.15. Model: stop time = number crep clouds. n = 20 nights.
(Proc. Stepwise, SAS 1982.) "Crep" is the number of minutes in 1 crep.
but they also stop earlier (Table 3), so that duration is not
significantly affected. Nor was the effect of crep on duration
significant. The major factor accounting for variation in duration was
male density (Table 4, Fig. 15). As shown above, the number of males
affects duration by affecting start time, not stop time.
The data above provide a measure of what time of day ANY males are
making signaling flights. This section shows how, within that time
period, the numbers of active males are distributed. The minute-by-
minute counts of males crossing the transect (made more or less weekly,
n = 24 counts) were summed over the 1983 season to produce Fig. 16.
Note that the graph shows the number of active males counted per minute
after sunset (MAS). As also noted by Lloyd (1966, 1979b) and Adams
(1981), the pattern shown in Fig. 16 is the same as that observed on
individual nights: male activity rapidly plateaus and then ends
Similar data were collected in 1984, but in crep units rather than
in MAS. For comparison, the data in Fig. 16 were used to approximate
the temporal distribution of 1983 males in crep units. Data from both
sources were used to produce Fig. 17. The comparison shows that the
distribution is fundamentally the same from one year to the next. These
data were equally weighted and combined to produce Fig. 18. These
frequency distributions (Figs. 16 and 18) will later be compared to
similar data on available females.
Timing of female availability
When are females sexually available, i.e., outside of their burrows
and responsive to male signals? To answer this question, individual
Table 4. Sources of variation in collustrans male duration.
Variable F value Prob. >F
Rsquare = 0.72
Crep did not meet 0.15 significance level for entry into the model.
Clouds did not meet 0.15 significance level for entry into the model.
number 0.01 (Duration increases '1 min per additional
Regression on dependent variable duration. Model: duration = number *
crep clouds, significance level for entry into the model = 0.15, n =
20 nights. (Proc. Stepwise, SAS 1982.)
females were tracked throughout their adult lives. Figure 19 shows the
distribution of available-female-minutes in 1983. Females were
available from 12 to 33 MAS, with a peak in number available at 22 MAS.
Figure 20 shows the distribution of available-female-minutes in 1982.
Females were available from 6 to 52 MAS, and some of these females
stayed out past 60 MAS (not shown). Peak numbers were available at 24
MAS. These same data were converted to crep units to produce Figs. 21
There is an obvious difference between 1983 and 1982 frequency
distributions. In 1983, females did not stay available as late in the
evening as females in 1982. The reason for this difference will be
Keeping track of individual females night after night produced data
on how, if unmated, females changed the timing of their availability on
subsequent nights. (If mated, there were usually no subsequent
appearances.) Females in 1982 went for as long as 10 nights without
mating. Figure 23 shows that these females came out earlier on
successive nights, a phenomenon predicted by Lloyd (1979b). Such
behavior is likely to increase the duration of overlap with male
activity. However, this strategy is not without accompanying risks (see
chapter on risk of predation).
Comparison of male and female timing
In this section the timing of female availability is compared with
that of male mate-locating activity. The female availability data are
presented in both MAS and crep units for each year they were studied,
1982 and 1983 (presented above in Fig. 19 through 22). The male
activity data taken in MAS during 1983 (Fig. 16) are compared with MAS
female data, and the combined male data in crep units (Fig. 18) are
compared with the female crep data.
1983 females. The comparison of 1983 males with 1983 female data
in MAS (Fig. 24) shows (1) male activity began slightly before the start
of female availability, (2) the time of greatest male activity
corresponded generally with the time of greatest female availability,
and (3) the number of females available dropped to 0 slightly before the
end of male activity. This is more-or-less the case predicted in Walker
(1983a) for a situation where females eclose over a 24-hour period, but
all become available to mate at a certain time of day. Such appears to
be the case with collustrans females. This 1983 female curve shows what
happens when males are abundant. Most females mated the first night
out, within a few minutes of leaving their burrows (see Lloyd 1979b).
The pattern is similar when the female data are converted to crep units
and compared to male availability data in creps (Fig. 25).
1982 females. Quite a different picture emerges from the
comparison of 1982 female availability with male activity. Figure 26
shows the MAS comparison with 1983 male data. Once again, the start of
male activity preceded the occurrence of female availability. However,
the number of males active declined rapidly at about the time female
availability was at its peak. In other words, a large proportion of
female availability occurred long after male activity ceased. The same
basic pattern is seen in the crep comparison (Fig. 27)
On 87 occasions females stayed out after male activity ceased
(1982), yet no mating was found to be initiated after males stopped for
the night. This does not completely rule out the possibility that males
do on rare occasions find females after male signaling flights end.
(For example, a male that perched for the night might walk to a female
glowing nearby.) However, it does seem reasonable to pursue other
explanations for why females were available when males were apparently
Although not unique (see Roberts 1971, Rutowski and Alcock 1980),
this situation seems to be a paradox from either the male or female
point of view. One would generally not predict that males would cease
their activity with so many potential mates available. Nor would one
predict that females would stay out after male activity ceased,
especially if it is dangerous to do so (chapter on risk of predation).
The timing of the end of male activity may be the easiest to
understand. Regardless of the number of females available, males
probably stop searching when the available light drops to a certain
level (Lloyd 1966; Adams 1981, 1982). Finding that males stop earlier
when it gets dark earlier (Table 3) supports this notion. The eyes of
collustrans males are adapted to searching in the failing light after
sunset (Lall et al. 1980). It may be that as the ambient light level
drops, collustrans males become unable to see and avoid obstacles in the
flight path (see Lloyd 1979b).6 Male eyes may be specialized to
maximize perception at the beginning of the activity period when the
number of potential mates is highest (see Lall et al. 1980). As seen in
Fig. 24 through 27, early evening males unable to search later have
potential mates regardless of male density. However, a hypothetical
male that searches later but is unable to search in the early evening
6If collustrans males are unable to see obstacles or predators (Lloyd
and Wing 1983) in darkness, this explains why they could not search for
females after the signaling period as some crickets do (Walker 1983b).
would have mates only when early evening males are scarce (also see
Dreisig 1971). If male eyes are adapted either to late or early
evening, but not both, then it is not surprising that collustrans males
would have evolved to search only in the early evening.7'8
Why did females in 1982 remain out after male activity had ceased?9
Females became unreceptive and re-entered their burrows as soon as
mated. But, when not mated, it seems that females "don't know when to
quit." That is, females do not seem to recognize that it is no longer
profitable to continue mating efforts due to a lack of active males.
If such fruitless extensions are maladaptive why don't females use
the same cues as males to time their activity or merely stop when they
no longer see males signaling? First, remember that females generally
get into this situation (males have stopped, the unmated female is still
out) only when males are scarce. $o if a female does not see any males,
it could be either because no males are in her vicinity (although males
are still active), or because males have stopped for the night.
Photinus tanytoxus is a sibling species of collustans that becomes
active as collustrans males stop for the evening. It would be
interesting to know whether speciation in this case was connected to the
visual specialization of tanytoxus for later and collustrans for earlier
flights (Lloyd, pers. comm.).
8Lloyd (1979) suggests that males use the flashes of other males as an
indicator of where to search, and a male that continues searching after
the others stop may stray from the area. Because males only search for
about 18 min per night, a lost male would have very little opportunity
to locate a deme by seeking flashing males. If this is the case, gene
flow between demes, which is already presumed to be limited because
females are flightless, may occur only rarely. Because even seemingly
trivial genetic differences can have major biological implications (see
for example Wing et al. 1985), the differences between demes should be
investigated (see Gross 1984).
That these were indeed collustrans females seems clear because many
mated with collustrans males on subsequent nights.
Therefore females would not be selected to retire in the absence of male
signals. On the other hand, females might be lured into continuing to
wait for a mate by the signals of a heterospecific male. Female
collustrans apparently cannot discriminate signals of their own males
from those of other species. Lloyd (1966) found that a collustrans
female he moved to a P. umbratus site repeatedly answered umbratus
males. I have observed collustrans females (n = 2) doing this at the
study site, where umbratus also occur at the same time collustrans males
are active. Lloyd also observed collustrans females answering P.
tanytoxus males. I took 2 collustrans females that were out late to a
nearby tanytoxus site, and they repeatedly answered males. Furthermore,
5 collustrans females at the study site would answer the first flash of
almost any flash pattern made with a dim penlight, although they varied
in how many signals they would respond to. One female even appeared to
respond to the light of a passing motorcycle. Therefore, females seem
unable to base a decision on retiring for the evening on either flashes
they do see or flashes they do not see.
With respect to using ambient light cues to time retiring, females
cannot rely on their micro-local light level to predict the levels their
males experience. Males fly in a relatively open, homogenous habitat in
terms of ambient light. Apparently, when it gets too dark to see
obstacles, they quit. The flightless females, on the other hand, are
perched on or near the soil surface. Some females are found on bare
sand, others in grass of varying height and density, and some beneath
the thick foliage of shrubs. The light level varies considerably from
place to place in these microhabitats where females (presumably with the
same ability to judge and react to light levels) are found. Therefore,
light level as a cue for when it is no longer advantageous to seek mates
is not nearly so reliable for females as it is for males.10
The second factor in females staying out too late is that the loss
of fecundity due to waiting another 24 hours to mate is a certainty (see
chapter on energetic costs), whereas the risk in staying out too late
(predation) is relatively low. Thus, if there is any doubt that males
have stopped, a female might be better off to delay retiring.
To summarize, the degree of difficulty experienced in finding a
mate varies considerably among collustrans females. Usually collustrans
males greatly outnumber males of the few other species that are active
at the same time of day. The collustrans female leaves her burrow in
the midst of a dense cloud of searching males. She has but to flash a
few times and she will be mated within seconds.
It is usually only during times when males are scarce that females
are still seeking mates at the end of the male activity period. In this
situation, females are caught between forces that dictate retiring and
those that dictate staying out. At this time I am unable to suggest a
better explanation than that females are without reliable information on
which to base a decision. However, that a possible explanation has been
offered should not close the issue. On the contrary, the question of
why females are available when males are not active should be vigorously
pursued. Anomalous phenomena, such as this, are of the greatest value
in science, for they show where our models are inadequate.
1Females may not use male flashes to cue when to leave their burrows
either. On an unseasonably cold night in May males did not make
signaling flights. I walked around with my penlight, and found 10
females. These were already perched outside of their burrows not coming
out, when they answered. This is another example of females being
available when no males were active (also see Dreisig 1971).
0 Females n=299
Apr May Jun Jul Aug Sep Oct
Fig. 14. Seasonal distribution of Photinus collustrans adults.
a) 1982. Number of new females found per week (7 days per week search).
1000 0 Females n=215 60
Males n=7359 4c
O -50 w
x -40 m
u- 600 \
0 .: 4- U
: : -300
200 I.*1 0
20 -n -rr ,p 0
Apr May Jun Jul Aug Sep Oct
Fig. 14. Continued. b) 1983. Line: number of males crossing 50 ft
transect per night. Bar: number of new females found per week (6 day
per week search).
0 Females n=70
k J.. kf
Apr May Jun
Fig. 14. Continued. c) 1984.
ft transect per night. Bars: Number
days per week search).
Males n= lLi (n
*_ _,, r...Ro.O-o- 0
Jul Aug Sep Oct
Line: number of males crossing 20
of new females found per week (6
* Males n=270
....-,. ....- -*,. ..' ". 0.S..
Apr May Jun
Aug Sep Oct
Fig. 14. Continued. d) 1985. Number of males counted or estimated
at entire collustrans site.
m A n~'
o 10 Females n=605
2 0.00 -
Apr May Jun Jul Aug Sep Oct
Fig. 14. Continued. e) Summation of data in Fig. 14a-d as
proportion of seasonal total. Line: males Bar: females.
0 0 0
Number of Males
Fig. 15. The relationship between density of Photinus collustrans
males and the duration of their activity period.
DoasueJl, 6uTsso3D saeW j o aaqtunN
86T 6uTssoaD sa'T W O3 aqmnN
V86T 6uTsso.D saTBW Jo laqumIn
W 2 CS a a ON r.
go~~c ~I 0P
~utsoi~ se~w ~O.L;o
$%5~88 3800 0
pumoa s8o1-Em3 ;o :xaq=N
-- ~ e <
n0 0 C
cc (s IV to J
punoa s I~mua JO JacqmN
I -- ----------- -- --------- g
_ ____^________ H *
punoa saTimaj jo aequmN
Fmnale Age (Day)
Fig. 23. Relationship of female start time to age.
number of observations at those coordinates.
0 \ 02
o0 \ 02 2
S1t 01 01
(E861) puno. saleinea jo aaqmunN
(E86T) BuTssoao saew ;o J aquinN
punoa sTeiurea o eqUnNN
, I In I 5
I~j ----- -
uccsso- s-ew ) IV OTou *2 E
(Z86T) punoA sOTeuIrJ jo .xaqunN
[ 8 8 t
I I I I
1---I ---- ---- ------- ----- Q l
I I 8 .
iM0 gii 0
6uTssO2Z3 sOeW sOTeW JO :LeqmnN
'. ^: J
-a>a ^ ^
_^^^ S 3
^yl" ------- ^ ^
^s^^^ --i un
UL \0 *
-- ---- ,T ^ -
BUTSOXO a"ewsa"BW30.laum '
punod saiemad Jo laqumN
co to :B D
I I *(D
"1 4 $
0 go t
u* Oo o
-' 1--------*S S in
SUMMARY AND CONCLUSIONS
Aspects of the reproductive ecology of Photinus collustrans are
documented in these studies and the adaptive significance of these
characteristics is investigated. The soft bodied, flightless females of
this species spend most of their adult lives in their burrows in the
soil. Females experience an increased risk of predation when outside of
their burrows. For this reason, adaptations that allow females to
minimize the time spent outside of their burrows are not unexpected.
Females leave their burrows only for mating, and even then remain near
the burrow opening. The vast majority of females mate only once, and
copulation durations are brief by firefly standards. On their first
night out, female exits from their burrows correspond with the peak of
nightly male activity. Thus, a female may mate and return to the burrow
within moments of her first exit, never to leave the burrow again.
Risk of predation is not the only cost of mating for females. A
female that fails to mate must delay oviposition to accommodate mating
efforts on subsequent nights. The longer the delay, the fewer eggs will
remain to oviposit. Unmated females make subsequent exits from their
burrows earlier, overlapping their availability with more of the nightly
male activity period.
Females that are unmated at the end of the male activity period
often remain out long after male activity has ceased. In so doing
females incur additional risk of predation while apparently having no
chance of mating. This unexpected phenomenon has not been conclusively
explained, and deserves further study.
Further study is also required to determine the cause of earlier
male start times with increasing density. Observations of other males'
signals could be a factor. Flashes of males flying on the previous
evening and/or "warm up" flashes by perched males before the onset of
the search period may be involved.
EPILOGUE: COLLUSTRANS IN PERSPECTIVE
The purpose of this section is to summarize characteristics of
collustrans' life history strategy and to make comparisons with other
species. Among Lampyridae females of some species, including
collustrans, are brachypterous, but other life history strategies of
adult females occur as well (Lloyd 1966). These include winged
flight-capable females, and apterous larviform females. This variety of
life history strategies within the family poses questions for
taxonomists (Cicero 1984) and behavioral ecologists (Fritz et al. 1982,
Gross 1984) alike. In order to address questions regarding how and why
the variety of female forms evolved, a data base concerning the
behavioral ecologies of the different forms is required (Cicero 1984).
Although beyond the scope of this project, an extensive survey of
species representing each of the three types of adult female body plan
alatee, brachypterous, larviform) would be valuable. Six
characteristics that should be considered are:
1. Position on semelparity-iteroparity scale.
2. Adult mobility and related behaviors--dispersal, search for
resources, and escape from predators.
3. Feeding or other nutrient acquisition as adult.
4. Adult lifespan.
5. Internal male reproductive structures.
6. Mating behavior.
There is a web of interrelationships among these characteristics, of
which the egg laying pattern (semelparity vs. iteroparity) may be the
key (see Fritz et al. 1982, and refs.). The criterion used by Fritz et
al. (1982) "in defining semelparity and iteroparity in insects is the
temporal pattern of egg maturation and deposition."
Almost by definition, iteroparity, which involves distributing egg
maturation and deposition over time and space, requires mobility. To
distribute eggs in space and/or to acquire nutritional supplies needed
to extend the lifespan for distributing them over time would almost
always require female mobility, although there are exceptions.
Females could acquire nutritional supplies by locating food and
feeding, but nutrients might also be delivered to them by courting
males. Such deliveries might take the form of nuptial feeding
(Thornhill 1980) or nutritious ejaculates (Bowen et al. 1984). Among
the fireflies nuptial feeding is unknown. As an indicator of the
possibility of nutrients being delivered in ejaculates, the complexity
of the male accessory glands is considered here. Little is known about
the function of firefly accessory gland products, but such products in
other insects are known to serve as nutrition (Boggs and Gilbert 1979).
Other functions are also known in other insects (Parker 1970). Related
to transfer of nutritious ejaculates is mating behavior, particularly
copulation duration and female mating frequency. In fireflies the
briefest known copulations apparently involve only semen transfer,
whereas longer copulations involve transfer of complex ejaculates (Wing
1985). Female multiple mating would be expected when females acquire
needed materials in the ejaculate, although it may also occur for other
reasons (W.F. Walker 1980, also see Watanabe 1988).
Using these six characteristics here to compare species
representative of each female body type will more clearly define the
Materials and Methods
Photinus umbratus, Pyropyga minute, and Photuris spp. were studied
at the collustrans site. Photinus marginellus and Phausis reticulata
were studied at Highlands Biological Station in Highlands, N.C. P.
marginellus was found in the yard at Illges Cottage and nearby, and
Phausis reticulata was found on the Rhododendron Trail.
Photinus marginellus and umbratus females are alate, and the same
methods of study were used for both species. Perched females were
located by their flashes in response to signals of flying males or to
penlight flashes. The locations of females were marked with numbered
stakes, and pair formations were observed. Copulation durations were
determined by intermittent observations in the field. Males and females
were dissected using standard techniques. Larvae were housed and fed in
the same manner previously described for collustrans.
Data on longevity of alate females are drawn from records of the
interval between capture and death or dissection of field-collected
specimens of unknown age. These females were kept in vials with a cube
of apple in the laboratory.
The larviform females of Phausis reticulata were located by their
glows. Female locations were marked with numbered stakes. Some females
were collected and kept in vials with soil and leaf litter from the
1Also at this site numerous phengodids are found when the area is
flooded (Wing 1984). Additional studies of Photuris spp. were made at
the nearby macdermotti site (Wing 1982).
site. Males were netted as they flew glowing through the forest, and
were housed in vials with damp leaves. Matings were arranged by placing
a male in the vial with a glowing female. Some males were dissected.
Results and Discussion
Species with Brachypterous Females: Photinus collustrans
In considering the subject of iteroparity vs. semelparity, three
questions should be addressed (Fritz et al. 1982). First, are eggs
deposited in different places? Second, are the eggs deposited at
different times? Third, are eggs matured at different times?
As has previously been discussed for collustrans, the eggs are all
deposited in one place (the burrow) and at one time, a single
oviposition bout lasting a few days.2 Evidence from 15 dissections
indicates that all eggs are matured at once. In five of the females a
few of the eggs were not yet quite full sized, but in no case were eggs
found in early stages of development. (As noted in chapter on energetic
costs, mature egg count was x = 71, range = 30 to 112, SD = 23.)
Other characteristics also suggest semelparity. While collustrans
females are capable of some movement, they use this ability only in
brief exits barely outside of the burrow for mating. No evidence of
feeding by collustrans females has been found. They are not equipped to
hunt underground as their larvae do, or to search for food outside the
burrow. Male collustrans deliver only semen to females in their very
brief matings and male reproductive systems are correspondingly simple
2The small percentage of females that do return for a second mating a
few days after the first probably deposit part of the complement of eggs
after the first mating, and the rest after the second. One female that
was dissected after returning for a second mating contained only 18 eggs.
in structure (Wing 1985). Furthermore, the vast majority of females
mate only once. Thus, it seems unlikely that the ejaculate is a
significant source of nutrition for collustrans females.
All of these characteristics support the classification of
collustrans as semelparous.
Species with Winged Females
The representatives of species with winged females for this
comparison are a congener of collustrans, Photinus macdermotti, and 2
Photuris species. Photinus macdermotti will be considered first.
Other alate firefly females have not been monitored 24 hours a day
for life in the field to determine oviposition patterns (reviewed in
Buschman 1977), and the same is true of Photinus macdermotti females.
However, on the question of egg maturation there are some data. Some
females, presumably recently closed, contained no mature eggs or
oocytes in intermediate stages of development. Other, presumably older
females, contained all stages. In each of 8 females, eggs in all stages
of development were found, and there was a wide range in number of
mature eggs (range = 10 to 93, x = 34, SD = 28). This may reflect
different ages or other circumstances among these field-collected
females.4 These data indicate that macdermotti females are capable of
maturing eggs over time, and so iteroparity relative to collustrans is
suggested. This is further supported by data on mobility.
I have not surveyed and analyzed published laboratory data, but some
are indicative of alate female iteroparity. For example, Buschman's
(1977) laboratory data indicate that some Pyractomena lucifera females
oviposit several clutches of eggs over a period of several weeks.
Photinus marginellus (n = 3) also had eggs in developing stages, which
may be characteristic of winged Photinus females in general.
Female macdermotti are known to move about in the field. Marked
females were found in different locations from night to night, and on
two occasions females apparently moved to where males were (Wing 1982).
By contrast collustrans females remained at their burrow night after
night, even when it was located under dense shrubbery where there was
almost no chance of being found by males. The mobility of macdermotti
females suggests that they are able to distribute eggs in space.
Female macdermotti probably acquire very little nutrition by
feeding. They apparently use their mobility to acquire water from the
underside of leaves (Wing 1982), and they could travel to other food
sources as well. However, no food was noted in the guts of dissected
females. (Food is obvious in the guts of predaceous fireflies, below.)
A more likely source of nutrients is the ejaculate of males. Male
macdermotti are characterized by complex male accessory glands (Wing
1985).5 As would be expected, these glands produce ejaculates that are
more complex than those of collustrans (Wing 1985). P. macdermotti
copulations are prolonged, at least in part, to accommodate the transfer
of the complex ejaculates (Wing 1985). These macdermotti couplings
last for hours, compared to about a minute for collustrans. Although
the functions of the macdermotti male products are unknown, and there
5Other species with alate females that share this trait include the
Photinus species marginellus and umbratus; Pteroptyx valida (Wing 1982,
Wing et al. 1983); and Pyropyga minute. A previous pilot study of
species from several genera (Lloyd and Dong Ngo unpublished) indicates
that as a general rule species with alate females have complex male
prolonged couplings also characterize other species with winged
females, including Pteroptyx valida (Wing et al. 1983), Photinus
marginellus (n = 2, '1.5 h and "1.75 h), and Photinus umbratus (n = 1,
>12 h but < 15 hr). Also see Dewsbury (1985) for a contrasting
comparison of rodent male products and copulation durations.
are certainly a variety of possible functions, one possibility is that
they are a source of nutrition to females. Consistent with this notion
is the fact that macdermotti and other alate Photinus females are much
more likely to mate repeatedly than are collustrans females (Wing 1985,
Sara Lawrence unpublished).
Photinus macdermotti females also seem to live longer than female
collustrans. Seven macdermotti females survived 2 to 4 weeks in the
laboratory after being collected at unknown ages. By contrast,
collustrans females survive only about 10 days (Wing 1982).
The characteristics of macdermotti females suggest that they have a
greater opportunity to mature eggs over time and to distribute them at
different times and places than do collustrans females. An even more
iteroparous example may be provided by the genus Photuris.
Photuris species of the versicolor-pennsylvanica group represent
extremely mobile forms of firefly females. This mobility is utilized
for hunting, which includes aerial predation of other fireflies (Lloyd
1984 and refs., Lloyd and Wing 1983). Thus, they acquire substantial
nutritional resources that could be used for egg production, and they
are capable of moving from place to place where they could oviposit.
Photuris are also extremely long lived by firefly standards. Four
Photuris "D" females survived 4 to 8 weeks after capture with no prey to
feed on. They are also apparently capable of maturing eggs over time.
Six Photuris versicolor females were dissected, and all were found to
7photinus umbratus females may represent an extreme in Photinus female
multiple mating. Females are usually found perched on the tips of grass
blades, facing upward. When investigating responses to penlight
signals, in 8 cases I found a female already in copula. In 4 other
cases the female was copulating and a second male was standing by. In
only 5 of the 17 cases did I find a solitary female.
have eggs in developing stages. (Number of mature eggs range = 10 to
45, x = 24, SD = 13.)
The extreme mobility, long lives, and rich diet of Photuris
strongly suggest that they are among the most iteroparous fireflies.
Species with Larviform Females
A pilot study of Phausis reticulata provided some data on a species
with larviform females. Nothing is known of egg maturation patterns in
Phausis reticulata, and the only data on the egg deposition pattern
comes from one captive female. She was housed in a zipper vial with
soil and moist leaf litter. During one 24-hr period she deposited a
clutch of eggs on the ventral side of her own abdomen (Fig. 28). She
was able to carry the eggs in this location using her legs anteriorly
and the tip of her abdomen posteriorly to move along with the abdomen
arched, preventing contact between the substrate and the eggs. By the
next day, only one egg remained attached to her. It is possible that
individual eggs from a clutch end up scattered in different locations,
one of the criteria for iteroparity. However, that eggs are produced in
a group rather than singly suggests semelparity. Whether more than one
clutch is produced is unknown.
Some observations were made to discern the extent of Phausis female
mobility. In an area that was searched nightly, the positions of two
Phausis females were marked the first night they appeared there. The
females remained stationary and glowing. Both were monitored until a
The asymetrical arrangement of light organs or "portholes" on Phausis
females may appear to be sufficient to identify individuals, so that
marking is unnecessary. However, observations of one female in
captivity indicate that glowing females do not always use all of the
same lights. Thus, the same female may show different patterns at
different times (see Lloyd 1965).
half hour after the last male was seen. Neither female appeared at
these marked positions on subsequent nights, nor were they found
elsewhere. No evidence of a burrow was discovered at the sites of any
of four females that were observed in the field. One of the four was
found on a tree trunk about 1/2 meter above the ground. Sivinski
(unpublished) watched a female move to the top of a log for mate
attraction. These observations indicate that Phausis females may be
somewhat more mobile than are female collustrans.
Although it is not known whether Phausis reticulata females feed as
adults, the larviform condition suggests that they may continue to
use larval food sources (see Schwalb 1961). Data from one Phausis
female indicate an adult lifespan at least as long as that of
Phausis reticulata males have accessory glands that appear more
complex in structure than those of collustrans, though not as complex as
the macdermotti-like Photinus species (Fig. 29). What little is known
about mating behavior indicates that Phausis reticulata copulations are
of intermediate duration. A Phausis reticulata female in captivity
mated with a captive male for 5 min, 5 sec. A few days later, after
depositing eggs, the same female glowed and a different male was
introduced into her vial. Copulation lasted 3 min, 32 sec. At the same
location, Sivinski (unpublished) observed a copulation in the field that
lasted about 20 min.
Phausis reticulata couplings occurred with the male mounted on the
female's dorsum, as collustrans do. This may be the norm for flightless
female fireflies (see Wing 1985, Schwalb 1961).