Genetic elements of courtship inDrosophila: Mosaics and learning mutants
Genetic Elements of Courtship in Drosophila: Mosaics and Learning Mutants
Richard W . Siegel 0 1
Jeffrey C. Hall 0 1
Donald A. Gailey 0 1
C h a r a l a m b o s P. Kyriacou 0 1
0 This research has been supported m part by grants from the Lewis Fund to R.W.S., from the NIH to J.C.H. (GM-21473), and from the SERC (in the U.K.) to C.P.K. aDepartment of Biology, University of California at Los Angeles , Los Angeles , California 90024. 2Department of Biology Brandeis University , Waltham , Massachusetts 02254. 3Department of Psychology, Umverslty of Edinburgh , Edinburgh EH8 9TJ, Scotland , USA
1 INTRODUCTION COURTSHIP: CONTROLLING STIMULI AND NEURAL SUB- STRATES Courting Mosaics Courtship Stimuli and Their Neural Processing EXPERIENCE-DEPENDENT COURTSHIP Aftereffects Following Courtship of Young Males Aftereffects on Receptivity of Females Exposed to Courtship Song Aftereffects Following Courtship of Mated Females Courtship by "Males" That Are Mosaic for Learning Mutations Clocks, Mutants, and Conditioned Courtship SPECULATIVE QUESTIONS AND CONCLUSIONS Do Courtship Signals Inform or Manipulate? Is There a Selective Advantage for
Drosophila; courtship; learning; circadian
In recent years workers in several laboratories have initiated fresh
approaches to the study of courtship behavior in Drosophila melanogaster.
These investigations have focused particularly on genetic, physiological,
and neurological aspects of these complex behavioral interactions. Such
studies have revealed subtle changes in sexual behavior, caused by either
genetic mutations or the fly's previous experience. In this paper we review
this work and speculate as to the biological significance of these findings.
At times our discussions take on a distinctive sociobiological flavor.
There is a voluminous literature on courtship behavior in D.
melanogaster. The classical accounts have been reviewed several times
Ehrman, 1978; Spieth and Ringo, 1983; Ewing, 1983)
. It is sufficient for
our purposes to mention here only the highlights of the earlier descriptions
and experiments. When a sexually mature male and a mature (virgin)
female fly are brought together, the male orients toward the female, taps
her with his forelegs, and follows her closely if she moves. During
orientation or following, the male extends one wing or the other at different
moments and vibrates it to produce a courtship song
(see Ewing, 1977)
Later in the sequence, the male extends his proboscis and contacts the
female's genitalia. The next step is one or more attempted copulations
in which the male curls his abdomen anteriorly from a position close to
and behind the female. These actions on his part seem to be facilitated
by one of the female's responses to the male's courtship, in that her
general movement becomes slower; therefore, she ostensibly becomes a
better target for copulation attempts
(e.g., Cook, 1977; Markow and
Hanson, 1981; Tompkins et al., 1982)
. If such attempts are to be successful,
the female will part her wings and separate her genital plates, allowing
The female's response to courtship differs when a mature male is
paired with a fertilized female. He follows, extends and vibrates his wings,
and may even attempt copulation; but the female responses now include
what are called repelling behaviors
(see discussion by Ewing, 1983)
most conspicuous of these is extrusion of her ovipositor. Also, the
positions of a mated ~ wings and genital plates do not change to
facilitate copulation, and, under laboratory conditions at least, courtship
is very rarely consummated.
The foregoing behavioral sequences are, seemingly, exhibited by flies
without sexual or social experience, although such an assertion relies
primarily on ancedotal evidence. A corollary is that courtship appears,
at least superficially, to be rather invariant between pairs of wild-type
flies. One infers that these reproductive interactions are narrowly
constrained behavioral traits reflecting the genetically determined anatomy
and physiology of the fly's nervous system.
A classical view states that courtship serves the dual functions of
synchronizing sexual arousal, especially in terms of overcoming female
inhibitions and male aggressions
, and permitting the
careful mutual assessment of mates before pairing. Hence, courtship
behaviors are thought, in part, to indicate to each member of a courting pair
something about the " v i g o r " of the other as a potential breeder of
"quality" alleles into the next generation. Visual, auditory, tactile, and
chemical signals are exchanged and are integrated so as to provide information
on both the readiness and the suitability of each mate for procreation.
The utility of an elaborate courtship display for the species would be to
allow for the recognition of conspecific partners of optimal genotypes.
The two key features of this formulation are the informational content of
the courtship behavior and the capacities of the flies to make choices
among different partners in accord with selective advantages. The
avoidance of wasted physiological and material resources is also an important
Recent results reviewed here suggest a somewhat different view of
the function of courtship in Drosophila. Courtship may serve to
manipulate, rather than to inform, a given member of a courting pair. In this
view, the fly providing courtship s i g n a l s l s o m e of the most important of
which are chemosensory cues (also see articles by Tompkins and by
Jallon, this issue)--manipulates the behavior of the potential mating
partner and gains an advantage over other flies of the same sex. The
integration of signals, such as reproductive pheromones, by higher functions
of the central nervous system is thus likely to serve selfish interests. The
behavioral results of this sensory reception and information processing
may have little to do with mate selection.
We discuss the ideas revolving around manipulative components of
courtship partly in the context of the putative adaptive value of learning
and memory in courtship. That there are experience-dependent elements
in the reproductive behavior of these flies is introduced by our review of
the discovery that general learning and memory mutations perturb certain
components of courtship in Drosophila melanogaster.
COURTSHIP: CONTROLLING STIMULI AND NEURAL SUBSTRATES
Several investigators have studied the courtship of mosaic flies with
the aim of locating the sex-specific foci for each reproductive behavioral
(see reviews by Hall, 1982; Quinn and Greenspan, 1984)
experimental goal has been to identify tissues which must be genetically
male or female in order that a mosaic fly express a characteristic
sexspecific behavior. Sex mosaics can be readily produced and recognized
in D. m e l a n o g a s t e r
(reviewed by Hall et al., 1976)
. These start out as
normal XX female embryos but lose an X chromosome early in
development so that the mosaic has both male and female tissues. Employing
a wide array of such gynandromorphs, with many different male-female
tissue distributions, it has been found from tests of their courtship
performance with females that the male-typical behaviors--tapping,
following, and wing extension--depend solely on male brain tissue (with the
genotype of sense organs being irrelevant). A basically normal song
(which in a mosaic may or may not accompany wing extension), plus
attempted copulation, occurs only in mosaics whose thoracic ganglionic
tissues are all or part male. In order to elicit courtship from normal males,
mosaics must have a female abdomen or one that is partly of this genotype.
In tests of the same type of mosaics with males it has been found that
female brain tissue is necessary for receptivity to the male's copulation
The influence of tissues on or in the female's abdomen, with respect
to the elicitation of courtship, seems consistent with the role that a variety
of chemical cues plays in the initial stimulation and later modulation of
male-female courtship interactions (again, see accompanying reviews).
It is likely that the female controls the manufacture or release of
pheromone-like cues from her abdominal tissues. Whereas olfactory cues,
whether from the female's abdomen or elsewhere, facilitate courtship and
mating, they are apparently not essential, in that mutants which cannot
respond to certain odors can mate (see Tompkins, this issue, for details).
Similarly, vision is useful but not essential since blind flies or those in
the dark can mate
(e.g., Kyriacou, 1981; Tompkins et al., 1982)
The results from these mosaic studies are significant in several
respects. First, they indicate the anatomical regions which must be
genetically of one sex or the other for courtship sequences to occur. Second,
they suggest that there is an interplay between sex-specific foci both
within a fly and between members of a pair of flies during courtship
and Benzer, 1976)
. Thus, the posterior parts of the female (the abdomen)
provoke the male to perform the earlier steps in courtship display, and
this in turn leads to receptive behaviors from the female. As noted above,
a fly must have male brain tissue if it is to express the male-specific
behaviors relatively early in the courtship sequence. Therefore, it can be
suggested that signals from the male head somehow trigger the thoracic
ganglia (when they are at least in part haplo-X) to mediate the later step,
attempted copulation. We elaborate further on these points in the
discussion that follows.
Courtship Stimuli and Their Neural Processing
Courtship behavior seems to begin as a male response to chemical
cues released constitutively by another fly. That is, it appears that
chemicals which stimulate male courtship (see Tompkins and Jallon, this issue)
do not depend on the presence of another fly for release. Thus, isolated
immature males and females that have been maintained separately from
males after eclosion liberate courtship stimulants (Tompkins et al., 1980;
Gailey et al., 1982). Females from a variety of other insect species depend
only on environmental changes in, for example, light and temperature to
release chemicals which attract males (see Leonard el al., 1974).
Similarly, there are no results in Drosophila which suggest that mature males
bring about or induce the release of substances which stimulate the onset
With respect to the behavior of the male toward females, it should
be stressed at once that all the major courtship elements occur in most
normal male flies which express any courtship at all
(see, for example,
Tompkins and Hall, 1983)
. Since that sequence of behavioral components
is independent of the total amount of courtship behavior performed, and
also of the so-called "attractiveness" of the female, it can be inferred
that the mechanism for its control is restricted to the male. Results given
next support this inference and have been chosen for review because they
have begun to suggest questions concerning the neural mechanisms
underlying these actions. A priori, one can imagine that there is a
head-tothoracic ganglionic flow of information which controls courtship motor
outputs (again, as inferred from the aforementioned mosaic analyses of
male-specific behavior). It is useful to consider this pathway in the context
of evidence from several neurogenetic studies in Drosophila
by Hall, 1982)
which suggest that much chemosensory and visual input
is processed in the brain and in regions of the head ganglia which are
becoming increasingly well defined.
Less work has been directed toward analysis of female-specific
courtship behaviors in D. melanogaster. However, the data which are available
reveal that the brain is again the primary site for input and processing of
male-generated '~ at least for reception of the auditory
component by the female's antennae
(reviewed by Ewing, 1977)
. The neural
substrates for this processing would again be "built into" the female brain
by the action of the genome. One of the female courtship responses is a
change in locomotor activity. Mature virgin females become quiescent
(e.g., Markow and Hanson, 1981)
for periods whose durations are
inversely related to the mating efficiency of the male (Tompkins et al.,
1982). Young males, 1 day old or less, include individuals which court
but cannot mate [see
Jallon and Hotta (1979)
and Tompkins et al. (1980),
who note that many such immature males do not court at all]. Mature
virgin females that are courted by such males do not reduce their
locomotor activity, nor do immature virgin females (1 day old or less) when
they are courted by mature males (D. A. Galley and R. W. Siegel,
unpublished). The female stopping behavior is evidently a response to
airborne chemicals released by the courting male, for this behavior is
defective in mature virgin females which are genetically s m e l l b l i n d
(Tompkins et al., 1982)
. Song-like auditory input, by itself, can also elicit
a slowdown in female movements
It is not surprising that these female-specific behaviors, or at least
overt receptivity to mating attempts, depend on a diplo-X brain
and Hall, 1983)
. One imagines that such receptivity is in part controlled
by the brain's processing of olfactory
(Tompkins et al., 1982)
input, in the same general manner that these categories
of stimuli are important to initiate processing events in a courting male.
In the case of mosaics analyzed for female receptivity, the hypothetical
processing centers for the inputs just noted could be the two dorsal,
anterior sites which are the minimal internal tissues that must be
diploX in order that a gynandromorph be receptive to male courtship
(Tompkins and Hall, 1983)
. In contrast, the requirement for male courtship, in
its early stages, is that cells in the dorsal posterior brain be haplo-X
1979; Schilcher and Hall, 1979)
Certain gynandromorphs are able to express both male and female
(Cook, 1978; Szabad and Fajszi, 1982)
. These mosaics
probably had mixed haplo-X//diplo-X brains, but no internal marking was
included in these studies. Other matters concerning the "genotypic
control" of male and female reproductive behaviors are discussed in the
following section, which describes some of the genes that are involved
in certain elements of courtship.
E X P E R I E N C E - D E P E N D E N T C O U R T S H I P
It is reasonable to suppose that the energy cost of courtship behavior
is offset by the benefits of sexual reproduction. Thus it is surprising that
wild-type males will avidly court conspecifics with whom they cannot
copulate. Courtship of sexually immature males and of fertilized females
would appear to be biologically inappropriate as well as wasted effort.
One response of these courted males and females is that they are
apparently compelled to exert themselves by performing repelling behavior.
Such fruitless courtship occurs almost every time a mature male is placed
with an immature male or with a fertilized female
(e.g., Gailey et al.,
Aftereffects Following Courtship of Young Males
For several hours after males emerge from pupal cases they neither
court nor copulate, and yet they have the surprising capacity to stimulate
mature males to courtship responses (see above). The mature male
follows, extends and vibrates his wings, then protrudes his probiscis at, and
attempts to copulate with, the young male. Whether wing vibration
produces a courtship song identical to that produced with mature virgin
(cf. Ewing, 1977)
is not known. The response of the young male
is to flick the wings and to decamp
(e.g., Cook and Cook, 1975; Jallon
and Hotta, 1979)
. Subsequent results showed that the courtship response
of the mature male decays and is soon absent (Gailey et al., 1982). When
20 mature males were individually paired with 20 immature males, all of
the mature males courted at the outset, but 30 min later only 10 expressed
any courtship. Moreover, the courtship index (CI; the percentage of an
observation period during which courtship is expressed) declined from
67 for the initial 10-min interval to 17 for the third 10-min period..
Further evidence for modification of coursthip behavior following
pairing with an immature male has come from observations of mature
males either kept in isolation for 30 min or paired with another m a t u r e
male for this amount of time. Males having undergone either of these two
types of experience courted immobilized young males with the high
intensity which is characteristic of the response to virgin females (CI, at
least 30). In contrast, males which first courted i m m a t u r e males
subsequently expressed almost no response to the immobile males (CI -~ 1).
The depressed response of mature males, having undergone this prior
experience, is retained for 4-5 h (Gailey et al., 1982).
These experiments lead to some thoughts about the mechanism for
this behavioral modification. First, it should be noted that this particular
example of experience-dependent courtship is "specific," in the sense
that males "trained" to no longer court immature males still court females
at normal levels (Gailey et al., 1982; 1984). Second, the phenomenon
cannot be explained as a general reaction to the overt rejection
experienced by the male, to "discouragement," or to behavioral fatigue. For
example, consider the following situation. Males will court immature
unreceptive virgin females, who perform their usual repelling responses
(Connolly and Cook, 1973)
and fail to copulate
courting these kinds of females show no diminution of courtship activity
for at least 30 min (Gailey et al., 1984).
Immature males, as courtship objects, exert altogether different
effects on the subsequent behavior of the courter. These depressive
aftereffects are elicited even when the immature males are unable actively to
reject courtship because they have been entirely immobilized. Such a
courtee is still courted by a mature male, and this is followed by the
typical courtship modification (Gailey et al., 1982). Even apparent odor
traces from young males are sufficient to cause a depressive aftereffect
when males exposed to such substances are subsequently placed with
immature males. Pairs of mature males will also court each other when
placed in small chambers previously occupied by immature males,
implying that there are at least traces of courtship stimuli left behind in the
chambers. In the context of this implication, it is significant that a mature
male who was isolated in a chamber which previously contained immature
males thereafter expresses a depressed level of courtship when he
subsequently is tested with an actual young male (Gailey et al., 1982). These
results indicate that the decrease in responsiveness of mature males to
young males is caused by chemosensory factors released by young males
(see Jallon, this issue, for more details); these chemicals induce both the
initial courtship response and the subsequent courtship modification.
Perhaps, then, modified courtship is caused by the mature male
learning to avoid the chemical odorants released by young males. This
suggestion was examined further by testing males from six mutant strains,
previously shown by Quinn and colleagues
(see reviews by Quinn and
Greenspan, 1984; Tully, this issue)
to be defective in associating chemical
cues with electric shocks. Males from five of these six mutant lines failed
to modify their response to immature males as a consequence of a prior
courtship experience (Gailey et al., 1982). The sixth mutant, cabbage
(cab), is especially instructive, in that males do exhibit courtship
modification, but the refractory p e r i o d - - w h i c h normally lasts 4-5 h - - i s
reduced to an hour or less (Gailey et al., 1982). These results indicate that
the mechanism for the modified courtship, which could be a case of
habituation (given the results on odor traces discussed above), may
involve several genetically specified steps which are also common to
Other aspects of nonassociative learning (outside the realm of
courtship behavior) are impaired by this same array of mutations
. Yet, the affects of these genes on "higher" and "simpler"
learning are not always the same, implying that the mechanisms share
some but not all of their genetically determined features. For example,
turnip (tur) flies [defective in associative learning by definition, because
of the tests used to isolate this mutant (Quinn et al., 1979)] are normal
in sensitization to sugar stimuli but are defective in habituation
and Quinn, 1982)
. The mutation amnesiac (amn) disrupts the later
retrieval of associatively learned information (Quinn et al., 1979); yet with
regard to courtship modification, amn blocks the initial capacity of males
to become conditioned by immature males (Gailey et al., 1982).
A simple mechanism which can account for all of the results of testing
mutant and normal males in these courtship situations is that immature
males constitutively release a chemical which stimulates mature males
to court them. The mature males soon become habituated to that
substance, decreasing their responsiveness in the presence of the first
immature male (Gailey et aI., 1982) and then being markedly unresponsive
to a second such male.
Aftereffects on Receptivity of Females Exposed to Courtship Song
Responsiveness of females to male courtship can in certain instances
be increased experimentally. The relevant phenomena seem to be in the
realm of behavioral sensitization. Here again the
associative/nonassociative conditioning mutants are pertinent. Females that had been
previously subjected to electronically generated courtship sounds
Schilcher, 1976a,b; Kyriacou and Hall, 1982)
begin to mate more rapidly
when they are mixed with males
(Schilcher, 1976b; Kyriacou and Hall,
. These enhancing aftereffects last about 1-3 min, in that the males
need to be introduced to the females within that time period (i.e.,
poststimulation) in order for the reduced copulation latencies to reveal
(Kyriacou and Hall, 1984)
. Superficially, at least, this phenomenon
would seem to be a case of sensitization, and it is interesting that several
learning mutations impair sensitization of the proboscis extension reflex,
in tests using sugar stimuli
(Duerr and Quinn, 1982)
, just as they impinge
(Duerr and Quinn, 1982)
. Three of these mutations, dunce,
rutabaga, and amnesiac, have now been shown to block or attenuate
(Kyriacou and Hall, 1984)
. Females homozygous
for either of the first two mutations show no aftereffects of song
prestimulation. Those heterozygous for either variant, or homozygous for
amnesiac, have shortened copulation latencies when mixed with males
just after prestimulation; but if one waits 1 min after the end of the song
stimulus and then introduces males to the prestimulated dnc/+, rut/+,
or amn/amn females, the enhancing aftereffects on copulation latency
(Kyriacou and Hall, 1984)
. Therefore the mating
behavior of these prestimulated females depends on their prior experience,
and the parallels between the effects of these mutations on associative
conditioning and those on acoustic priming are striking. Such similarities
extend to the fact that mutant/wild-type heterozygotes for dnc and rut
forget more quickly in both associative conditioning and in acoustic
(Dudai, 1983; Kyriacou and Hall, 1984)
Aftereffects Following Courtship of Mated Females
There is a demonstrable energy cost connected with the courtship
of a fertilized female. Males court such females with a reasonable degree
of intensity (though see below), and the mated females persistently
generate repelling behaviors
(e.g., Connolly and Cook, 1973)
. The frequency
of copulation is about 1% under laboratory conditions when the females
are tested 1 day after they initially mated
(e.g., Siegel and Hall, 1979)
Intuitively, courtship directed at mated females would be expected to be
relatively limited, and this is the case: males paired with fertilized females
(which had copulated one day previously) showed an approximately
twofold decrease in their amount of courtship behavior during a 60-rain
observation period; there was no decrease in courtship activity for males
paired with virgin females
(Siegel and Hall, 1979)
. Although the sex
behaviors of both the male and the fertilized female waned during these
observation periods, three further experiments served to focus attention
on the male. First, it was noticed that males which had courted fertilized
females thereafter expressed diminished levels of courtship activity with
virgin females. The male courtship responses to virgins were inversely
related to the length of time that these males were paired with fertilized
females. Second, when males which had experienced courtship with
fertilized females were tested with virgins, it was found that the duration,
or quantity, of each of the courtship components was reduced, but the
quality of each activity and their sequence was not disturbed. Third, such
modified (i.e., depressed) courtship behavior persisted for 2 to 3 h in
(Siegel and Hall, 1979)
. Fertilized females, however, express
no behavioral modification following their limited experience with a male;
these females repel subsequent males with the same intensity and
persistence as before (R. W. Siegel, unpublished).
This aftereffect on male behavior has been termed an example of
conditioned courtship, in that it seems to involve a specific
experiencedependent behavioral change. The alternative possibility, that fertilized
females have a generally depressing effect on male behavior, is excluded
by the following observations. After being paired with mated females,
males do not appear to be generally debilitated. Most interestingly, males
from the a m n e s i a c strain recover the normal response to females after
only 30 min
(Siegel and Hall, 1979)
. If some sort of debilitation is to
explain the aftereffect, then one would imagine that a m n should show as
lengthy an aftereffect as do normal males. Finally, males which have been
" t r a i n e d " by mated females (i.e., males which exhibit relatively weak
courtship at subsequent females) remain responsive to immature males
(Gailey et al., 1984).
Decrements in male courtship following pairing with fertilized
females are limited to interactions with females. Reduced courtship toward
females occurs only in those males which have previously courted
fertilized females, not for those which, for example, had been courting young
(Galley et al., 1984)
. Furthermore, D. A. Galley and R. W. Siegel
(unpublished) have shown that male courtship behavior can be
conditioned by fertilized females which were immobilized either by disruption
of their central nervous system (decapitation or crushing of the head) or
by the effects of a temperature-sensitive paralytic mutation (as long as
such immobile females had been courted first; see below). Hence,
elaborate rejection behaviors are not a requirement for the aftereffect. Nor
is it likely that female repelling behaviors are sufficient to bring about
conditioning. This was concluded from observations of immature virgin
females, who might reject a male for an hour without causing him to court
poorly during or after such an experience.
It is chemicals which are uniquely associated with fertilized females
that play a decisive role in this case of courtship behavioral modification.
Tompkins et al. (1983)
showed that olfactorily defective males cannot be
conditioned as are wild-type males. It had been previously demonstrated
that extracts of volatile compounds from fertilized females can induce
two mature males to court each other
(Tompkins and Hall, 1981)
from fertilized females also reduce the courtship that a male performs in
response to a live virgin female
(Tompkins and Hall, 1981)
that fertilized females are associated with an "inhibitory" chemical cue.
These observations also suggest that males simultaneously perceive both
courtship-stimulating and aversive chemical signals which are provided
by fertilized females (again, see Tompkins' accompanying review for
details). The male might associate these two stimuli with an object of
courtship, and he thus learns to avoid female courtship stimulating signals
later o n - - e v e n when they are unconnected with aversive cues, that is,
when he subsequently encounters a virgin female. The male would in a
sense be anticipating that this courtship object will emanate a
hypothetically ~'noxious" substance.
Evidence on the manner in which this apparent aversive material
is involved in conditioning the male's later courtship behavior comes
from experiments designed to reconstruct, somewhat artificially, the
combined cues necessary to effect training. A male isolated in a small
chamber, containing an extract of fertilized females along with another fly,
subsequently expresses courtship modification; extracts from virgin
females and from mature males have no such effect
(Tompkins et al., 1983)
The " o t h e r fly" in this situation may be a virgin female or even a paralyzed
(Tompkins et al., 1983)
. This courtship object must be paired
with the fertilized female extract simultaneously, not sequentially, in
order for an aftereffect on the male trainee to occur
(Tompkins et al.,
Conventional theories of learning would suggest that if males do
associate courtship-stimulating signals with the putative aversive signal,
then the aversive signal must be perceived after and not before the
stimulus provoking courtship. Evidence for such a temporal sequence would
offer additional support for the existence of an aversive cue in these
behavioral phenomena. When a fertilized female is immobilized after her
first mating but before she is courted again, and then is paired with a test
male, that male courts the female but does not express later courtship
modification. However, if the fertilized female is courted for a brief
period, then immobilized, a test male with which she is paired is conditioned.
(Jackson et al., 1983; Gailey et al., 1984)
fertilized females must be courted and must have normally functioning
central nervous systems at the time they are courted if they are to
condition male behavior.
The idea that fertilized females release an aversive conditioning signal
in response to a courting male has been examined further. Males from
the mutant strain Don Giovanni (dg) court fertilized females as if they
were persistently courting virgins. Afterward, these males express
vigorous courtship with test virgins (i.e., the males were not trained). These
dg males are conditioned in a normal manner by courtship with immature
males, and the general behavior of flies from this strain seems normal
(cf. Gailey et al., 1984). Mutant dg males can be conditioned by courtship
with fertilized females but only in special circumstances. If fertilized
females are courted by wild-type males, immobilized, and then courted
by dg males, the dg males show a normal decrease in their courtship with
virgins (i.e., as occurs with regard to wild-type males). In addition,
fertilized females which are courted by dg males prior to immobilization do
not condition the behavior of wild-type males (D. A. Gailey and R. W.
Siegel, submitted for publication). Hence, we infer that these mutant
males are defective in the capacity to induce fertilized females to release
the aversive signal; this would explain why dg males do not normally
express conditioning effects. If these ideas are correct, dg males should
be as capable as normal flies in other associative learning tasks. This was
found to be so. When artificial odors were presented together with electric
(cf. Quinn and Greenspan, 1984)
, males learned to avoid those
odors (R. W. Siegel and D. Byers, unpublished). Therefore the mutant
can express conditioned behavior in circumstances which bypass its
When coupled with an object of courtship, an extract prepared from
fertilized females will condition male behavior, whereas an extract from
virgin females will not. A male is thus initially required to fertilize the
female, who then becomes capable of eliciting a male's
experience-dependent behavior. Our observations indicate that copulation but not
fertilization is required. That is, females which have mated with sterile males
can condition other males (R. W. Siegel and J. C. Hall, unpublished),
whereas females successfully fertilized by mutant Don Giovanni males
do not condition (see Gailey et al., 1984). Male-specific e l e m e n t s - - s u c h
as the antiaphrodisiacs transmitted during copulation or the inhibitor made
in the female's reproductive tract from that antiaphrodisiac (see
Tompkins, this issue)--together with the courtship-stimulating cues which still
emenate from mated females are requisites for the male's behavioral
A series of observations on males derived from mutant strains which
are known to be defective in their general learning abilities lends additional
support to the notion that aftereffects caused by exposure of males to
fertilized females are in the realm of conditioned behavior. Gailey et al.
(1984) found that males of all the relevant strains are markedly defective
with respect to courtship conditioning, following pairing with fertilized
females. These were males expressing any of three different dnc mutations
or males from the cab, rut, or tur strain. Such mutants were shown to
be normal with respect to eliciting conditioning cues from fertilized
females (see above). That they are unable to respond normally to this
courtship training regime seems, again, to be due to faulty information
storage or retrieval. In one mutant, rut, the problem may be in storage
or even perception of mated females (i.e., the training fly). All the other
mutant males can tell the difference between mated and virgin females
(there is generally twice the amount of courtship with virgins as with
fertilized females); but rut males courted fertilized or virgin females with
equal and relatively high vigor (Gailey et al, 1984). Perhaps their olfaction
in this behavioral situation is mediocre. The rut gene is known to disrupt
(see review by Aceves-Pifia et al., 1983)
, an enzyme
defect which might impair behaviors other than learning.
Earlier observations by
Maynard Smith (1956
) showed that this
phenomenon is not limited to D. melanogaster. As part of a larger
investigation of mating behavior in D. subobscura, he reported that whereas 47
of 52 males without prior courtship experience promptly copulated with
virgin females, only 5 of 13 males which had courted fertilized females 3
h earlier copulated with virgin females. Seven of the remaining eight males
copulated with virgins after a day of isolation. In addition,
) reported that males kept in groups, rather than in isolation, court
inefficiently; this effect disappeared after males were isolated for a day.
A similar grouping effect was reported by
Bastock and Manning (1955)
for D. melanogaster and has been since confirmed
. Nothing is known of the mechanism involved.
Courtship by "Males" that Are Mosaic for Learning Mutations
Mosaic analyses of genotypic differences related to courtship have
recently moved beyond the matter of gynandromorphs per se. We have
begun to assess the conditioned courtship of flies that are phenotypically
male and are mosaic for normal and mutant alleles of various learning
mutations. Since the pattern of mosaicism within these males' nervous
systems can be revealed histochemically
(cf. Hall, 1979; Schilcher and
, we hope to identify "learning foci" for these mutations, which
could be interpreted as neural substrates for particular components of
experience-dependent courtship. We have no predisposition as to what
these neural substrates might be. For instance, might the mosaics reveal
that a given mutation must be expressed in only a very limited portion
of the brain in order that defective learning occur, or will this analysis
show that most or all of the CNS must be mutant if the fly's behavior is
to be abnormal?
Tests of conditioned courtship seem as if they will be an ideal way
to study learning and memory in a series of mosaics, because the flies
can (and must) be tested individually. A second requirement is that the
normal vs. mutant behavioral differences be essentially all-or-none. This
is fulfilled in the case of dnc-induced defects in conditioning of mature
males by exposure to young males (Gailey et al., 1982). That is, almost
every dnc + male learns (courts poorly after training), whereas nearly
every dnc male does not (continues to court vigorously). We have,
therefore, constructed a preliminary series of dne+//dnc mosaics. Each one
was part diplo-X (i.e., heterozygous for a dnc mutation) and part
haploX (i.e., hemizygous for this mutation). The basic courtship behavior of
such a sex-chromosomal mosaic is often nonexistent or extremely
aberrant (as was implicit in the earlier discussion of the mosaic mapping of
male-specific and female-specific neural foci). Thus, we turned each of
our dnc+//dnc mosaics into a pseudomale, by introducing homozygosity
for the transformer mutation
(one of the several sex determination
mutations in D. melanogaster; see Baker and Belote, 1983, for review)
Diplo-X flies which are tra/tra look like males (though not entirely) and
behave in a normal male fashion
(Hall, 1979; Schilcher and Hall, 1979;
Kyriacou and Hall, 1980)
Behavioral testing followed by histochemical analysis of our
internally marked dnc mosaics has led us to the tentative conclusion that the
focus for the primary action of this gene is in the male's brain (D. A.
Gailey and J. C. Hall, unpublished). It remains to be seen, from the
necessary study of several more mosaics of this type, whether this brain
focus will hold up and whether the focus in these tissues is narrowly local
There are already some implications of our preliminary results. For
example, the d n c - e n c o d e d gene product, which is almost certainly a
particular form .of cyclic AMP phosphodiesterase
(reviewed by Davis and
, is present in essentially all body segments of the fly,
although it is rather concentrated in the head
. Yet, our
results suggest that the presence of this enzyme with respect to learning
seems to be important only in the brain. Otherwise, our mosaics with
large amounts of mutant tissue in posterior regions would have learned
There is one more related neurochemical matter which can now be
Tempel et al. (1984)
have recently shown that
temperaturesensitive dopa-decarboxylase mutants ( D d # s) are defective in associative
conditioning, as measured in the aforementioned mass fly tests (also see
Tully, this issue). The authors also report that, in single-fly tests, D d c ts
males (raised at a low temperature, then heat treated as adults) are
aberrant in the aftereffects they exhibit subsequent to courtship of mated
females. Similar defects have recently been revealed in analogous tests
using immature males to train the D d c ts males (D. A. Gailey and J. C.
Hall, unpublished). The enzyme being manipulated here is, like the
phosphodiesterase, distributed fairly ubiquitously in the fly
(e.g., Wright, 1977;
Livingstone and Tempel, 1983)
; but a depletion of dopa-decarboxylase,
and the consequent lowering of dopamine and/or serotonin levels
Livingstone and Tempel, 1983)
, may affect conditioning by exerting quite
local effects on the relevant brain functioning. This possibility needs to
be addressed in tests of D d c + / / D d # s mosaics.
Another potential implication of experiments on mosaics expressing
a given behavioral/biochemical mutation is that they can lead to a
comparison of neural substrates underlying associative vs. nonassociative
learning. Our tests of d n c mosaics thus far have involved "young-male
learning," an experience-dependent effect which appears to be
nonassociative habituation. With regard to "mated-female learning," which
seems to be more in the realm of associative conditioning, there is now
a test which essentially creates the necessary all-or-none behavioral
results of training normal and mutant flies. In the initial report of
dncinduced effects on mated-female learning, Gailey et al. (1984) showed
that three mutant alleles of the dunce gene cause defects in courtship
conditioning. These tests were done with virgin females as the "tester"
flies, with whom the males were placed after previous exposure to mated
(cf. Siegel and Hall, 1979)
. The effects of training are not
extremely severe in terms of courtship decrements expresged in the presence
of such virgin females
(cf. Siegel and Hall, 1979)
. But when mated females
are used as testers (as well as trainers), then the aftereffects last at least
12 h (D. A. Galley, unpublished). Furthermore, every dnc § male
expresses a high level of training, and every dnc male shows poor training
that does not overlap with the performance of the normal males (S.
Kulkarni, unpublished). This background of behavioral results will make it
possible for us meaningfully to analyze a series of dnc+//dnc mosaics for
their ability to be associatively conditioned. In fact, it should be possible
to test a given mosaic for association-type conditioning with fertilized
females and then to reuse this mosaic in a test for nonassociative
habituation with young males. Some mosaics may be able to learn by one
mode but not the other, unlike normal males, who can store and retrieve
both kinds of information (D. A. Gailey, unpublished). If such results
were found in tests of the mosaics, they would imply that there might be
separable neural foci underlying these two apparently different types of
learning which, nonetheless, are connected formally by virtue of the fact
that several of the conditioning mutations affect each of these categories
What structures in the Drosophila brain might be important for
learning? When we proceed with the mosaic studies discussed above, we
should keep the mushroom bodies in mind. These dorsally located
calyces, peduncles, and lobes have been considered "association centers"
in many arthropods and are involved in learning in social insects such as
(see review by Menzel and Erber, 1978)
. Drosophila has mushroom
bodies, albeit rather meager ones (compared to bees), and there is now
a series of mushroom-body mutants available (Heisenberg, 1980). These
were isolated on anatomical criteria and are now being studied
behaviorally, especially with regard to a variety of olfactory and conditioning
tests. Two of the mutants have been shown to be defective in mass-fly
associative learning. One mutant which is abnormal in these tests is
mushroom-bodies deranged (mbd)
(M. Heisenberg and D. Byers, unpublished;
cf. Heisenberg, 1980)
. Another is mushroom-bodies-miniature (mbm)
Borst and M. Heisenberg, unpublished; see brief description of this
mutant by Technau, 1984; Quinn and Greenspan, 1984)
. This is a particularly
interesting case, because mbm females have very small mushroom bodies
and exhibit weak shock-odor learning; but mbm males are normal, both
anatomically and in this component of conditioned behavior.
With respect to courtship, M. Heisenberg (unpublished) has collected
preliminary data showing that rnbd males (which are structurally quite
abnormal) are somewhat defective in the habituation tests involving
immature males. It was perhaps disappointing, however, to find that the
same mutant males exhibit normal aftereffects subsequent to their
courtship of mated females
. Yet the other brain mutants
have not been tested in this way, nor have any of them been examined
in memory tests related to courtship modification. Whatever these
experiments eventually reveal, the mosaic studies of the learning and
memory mutants should determine whether the mushroom bodies or other
brain regions--or all of t h e m - - a r e involved in conditioned courtship.
Given the new and expanding information on the behavioral
significance of these brain regions, it is worth recalling our earlier conclusions
from mosaic studies of basic male behavior. The dorsal, posterior brain
focus, which must be haplo-X if a D. melanogaster gynandromorph is to
exhibit any male-like courtship actions, seems to be rather near the
cortical " K e n y o n cells" which innervate the peduncles of the mushroom
and Hefsenberg (1980) for relevant facts and
discussion]. It could be that the morphology and the function of these bodies
are sex specific and are components of the neural substrate Which controls
reproductive behavior. This general proposal is made by analogy to the
well-known sexual dimorphisms in the brains of various vertebrates and
their presumed role in sexually dimorphic behavior
(e.g., Arnold, 1980;
MacLusky and Naftolin, 1981)
. It is therefore of interest that the
mushroom-body peduncles in D. melanogaster are in fact sexually dimorphic
. This anatomical difference may be
Clocks, Mutants, and Conditioned Courtship
The response of an organism to environmental cues can be largely
dependent upon the temporal pattern in which the cues occur. There are
narrow limits within which the temporal properties of that pattern can
vary if associative or nonassociative conditioning is to occur. Given these
generalities, one might imagine that disruption of the internal timing
devices of an organism could also upset conditioning. Mutations which
disturb both daily (circadian) rhythms and the rapid cycling rhythm of
the courtship song are available
(reviewed by Konopka, 1984)
et al. (1983)
discovered that males carrying any one of five independently
isolated mutations previously shown to lengthen the period of the fly's
rhythms respond inefficiently to conditioning cues. Controls on these
experiments excluded artifacts, such as changes in sensory perceptions,
alterations in cues, time of day, and abnormal courtship behaviors per
se. Therefore, conditioned responses to immature males and fertilized
females were disturbed, suggesting that these mutants with their "slowly
ticking" internal timers are perturbed in mechanisms common to both
types of learning. Consistent with this metaphor, it seems, is the fact that
short-period per ~ males are normal in conditioned courtship
; but it is perhaps odd that arrhythmic p e r ~ flies are also normal
in these tests
(Jackson et al., 1983)
, as they are in their general learning
(Aceves-Pifia et al., 1983)
Can these results reveal anything about the mechanism basic to
conditioning? Even if they do not, they may be of some heuristic value.
Autonomous or endogenous rhythmic neuronal firing has been noted
among both invertebrates and vertebrates
(reviewed by Delcomyn, 1980)
So-called "bursting neurons" almost certainly exist in Drosophila. It
would not be surprising if their firing patterns were linked with internal
timers which are in part controlled by the action of clock genes. Could
it be that the number of neuronal bursts per time increment is involved
in information storage and retrieval? If so, one could imagine that, in
long-period mutants, the number of bursts in a time interval may be low
enough to upset conditioning. Or perhaps conditioning requires that
autonomous firing patterns need to be "in register" with the transmission
of signals from the environment. It will potentially be very informative
to study mosaics for the per-long and andante mutations
1984; Jackson et al., 1983)
, using courtship conditioning, as the behavioral
tests. Brain regions affected by these mutations, with respect to these
traits, could conceivably be the same as those affected by conditioning
mutations such as dnc. Could it also be that the foci influenced by these
clock genes, with regard to the various rhythm phenotypes
1984; Kyriacou and Hall, 1980)
, are neural tissues strongly related to
those which are affected by the learning and memory variants? An
analysis of mutant vs. normal rhythms in mosaics expressing a per mutation
has proven feasible (Konopka et al., 1983). This implies that a comparison
of the "clock foci" to the "learning foci" will eventually be possible as
SPECULATIVE QUESTIONS AND CONCLUSIONS
Do Courtship Signals Inform or Manipulate?
Courtship signals are sensory cues which are produced by one
organism, perceived by a conspecific, and are effective in evoking in the
latter the characteristic actions which normally (but not inevitably)
culminate in copulation. It is convenient to refer to the signal producer as
the " a c t o r " and the signal recipient as the " r e a c t o r . " It is implicit here
that courtship signals can be distinguished from other signals, such as
alarms, and that the former are in some way "specified by the genome,"
at least in the sense that they are to some degree species specific. As for
the function of courtship signals, it is a commonplace idea that they
typically serve to coordinate the behaviors and, in some cases, to bring
together prospective mates. None of the recent work on reproductive
behavior in Drosophila melanogaster appears to be in serious
disagreement with these statements. Courtship signals in this species, however,
can bring together and stimulate interactions between flies which cannot
copulate with one another. This deserves further comment.
If courtship signals are to be effective, they must modify the behavior
and physiology of the reactor, bringing it into functional accord with the
behavior of the actor. At this point, we pause to note that behavioral
biologists can perhaps be divided into two camps. One holds the seemingly
traditional view that courtship signals serve only to convey information
from the actor to the reactor. This is followed, presumably, by the
reactor's integration of the signals, producing behavioral changes which
ultimately result in copulation. Another view, which seems to be less
widely held or at least more recently invoked, is that there is little or no
informational content in the signals per se; instead, these signals are
means b y which the actor manipulates the reactor to his own advantage.
Reducing these contrasting views to their simplest terms, we ask whether
courtship signals in Drosophila inform, or do they instead persuade the
recipient? It should be mentioned, as well, that a classical ethological
view embodying information exchange stresses the mutual benefit to both
actor and reactor; such would enhance, perhaps, the evolution of the
communication system. Others, in contrast, argue that communication
benefits the actor only and that if there is any benefit to the reactor it is
at best incidental and probably irrelevant. The two viewpoints are stated
most forcefully by
Dawkins and Krebs (1978)
, and Caryl
The suggestion that courtship signals function to convey information
from the actor to the reactor presents several difficulties in considering
reproductive behavior in D. meIanogaster. For example, a particular
signal from one individual can convey what is true at one moment; at the
next, the signal lies. Thus, a sexually immature female, who repels a
male's copulation attempts and is unreceptive to them, nevertheless
produces courtship stimulating chemicals. If the cue is informational, its
information is false. A few hours later, this female is a mature virgin and
will copulate in response to male courtship. The same (or at least the
same general type of) courtship-stimulating cues are released as before,
but their informational content has in a sense changed from false to true.
Finally, fertilized or aged females continue to produce
courtship-stimulating materials; yet they do not normally copulate
(e.g., Connolly and
, so the stimulating signal once more contains false
information. Consider, too, that immature males induce mature males to court
them; this is difficult to reconcile with an hypothesis which holds that
signals have informational content of benefit to both actor and reactor.
Our knowledge of the detailed aspects of courtship signals in D.
meIanogaster favors the idea that some signals are manipulative,
especially in light of the ostensibly false stimulating cues that have just been
discussed. It would seem, for example, less energetically expensive to
both males and females if a fertilized (or an old female) were to shut off
production of courtship-stimulating substances. The fertilized female
possesses aversive substances as well as aphrodisiacs; the former do cause
relatively depressed responses on the part of the male. By provoking the
male to court, the mated female can manipulate him in order to benefit
herself as the actor. This is because a male which courts a fertilized
female is distracted from courtship of and copulation with a virgin female.
By eliciting a male to court, the mated female gains an advantage over
virgin females, because the probability that such virgins will be fertilized
and contribute to the next generation is diminished. Therefore, the
capacity of fertilized females to elicit courtship, in part through their
continued production or release of aphrodisiacs, provides a means of
manipulating male behavior to their advantage. It could also be argued that
immature and aged females distract the attention of males from other
females, to the benefit of any future progeny or those that are developing.
In summary, the behavioral interactions we have discussed can be
viewed in the context of courtship signals serving to manipulate rather
than inform. The female manipulates the male, causing him to fertilize
her eggs. Following fertilization she manipulates secondary males through
her signals in order to gain advantage over other females. The mating
male manipulates the female, causing her to help him in gaining an
advantage over secondary males. The manipulated male here may even gain
some benefit himself within this system. We now turn to a discussion of
this putative benefit and to others which are related to conditioned
behavior in these flies.
Is There a Selective Advantage for Drosophila's Ability to Learn and
The attractiveness of a fertilized female could provide the male who
mated with her a means of gaining an advantage over other males. The
male here provides as his legacy a female that can occupy the attentions
of his male competitors. Hence, his own chances of finding another virgin
female to inseminate could be improved. In addition, the aversive stimuli
associated with mated females may cause males subsequently to avoid
not only other mated females (which it is a good idea for them to avoid)
but also virgin females.
A male trained with a fertilized female for an hour takes 2-3 h to
rebound back to normal, whereas this aftereffect lasts between 12 and 24
h if the male is again paired with a mated female. Perhaps the initial
experience with the fertilized female subsequently biases the male's
responsiveness more toward virgins than mated females than would
otherwise have occurred (even though responsiveness in general is
depressed). In a situation where most females are fertilized, one imagines
that learning to avoid these females more strongly than virgins could in
effect sharpen up the male's discrimination between the two. There might
even be a selective advantage to males who modify their courtship in this
way. Such interpretations are rather counterintuitive. It of course appears
to be maladaptive that males can be conditioned with fertilized females
to show depressed responsiveness to virgins. However,
Hewitt et al.
have examined the genetic architecture for olfactory conditioning
in shock-avoidance tests, and their results have revealed a close
relationship between this type of learning and biological fitness. Since this
conditioning phenotype has common genetic elements with courtship
modification tests using both young males and fertilized females, we
assume that the aftereffects on male behavior of these courtship
experiences are not artifactual--in the general sense, and especially with
regard to depressed courtship of subsequent mated females.
Turning to the sex appeal of immature males, this attractiveness could
be viewed as a mechanism whereby newly emerged males decrease the
mating opportunities of older males and therefore increase their own
potential for mating when they have matured. An additional speculation
would be that the young male " w a n t s " to be courted in order to "sharpen
u p " his own courtship behavior for later on. Perhaps there is a critical
period during the development of the adult male when his eventual singing
behavior is susceptible to influence from a mature song. Such a
phenomenon would be analogous to the development of song in some birds
(Arnold, 1980; Konishi and Gurney, 1982)
. Such a situation could cut both
ways in terms of who is the manipulator and who is being manipulated.
The young male may require hearing a robust, species-specific song to
fully mature his own, whereas the older male may " c h e a t " and sing a
quite untypical song in order to disrupt this development. This (absurd?)
idea can be easily tested by comparing the songs of mature males directed
to virgin females with those sung to young males.
G e n o t y p e
W i l d t y p e
d u n c e
d n c / +
r u t a b a g a
r u t / +
c a b b a g e
t u r n i p
t u r / +
a m n e s i a c
D d d s
p e r i o d - l o n g
p e g
p e r ~
a n d a n t e
p s i - 2
p s i - 3
m b d
D e f e c t i v e E x p e r i e n c e - D e p e n d e n t
C o u r t s h i p in B e h a v i o r a l a n d N e u r o l o g i c a l
I t s
m e r e
o r y
S h o c k - o d o r
a s s o c i a t i v e
l e a r n i n g
i t s
m e r e
o r y
S u g a r - o d o r
a s s o c i a t i v e
l e a r n i n g
I t s
m e m
o r y
H a b i t u a t i o n
S e n s i t i z a
t i o n
( t o s u g a r s t i m u l i )
( - )
( - )
( - )
Conditioning mutants, clock mutants, biochemical mutants (some of which overlap the first two categories), and an anatomical brain mutant
have been tested m various features of the apparent learning and memory which accompany male or female reproductive behavior. The
clock mutants (per's, psi's, and andante) are described by
Jackson et al. (1983)
; all have long-period phenotypes except
the short-period pet ~ and the arrhythmic p e r ~ mutant. The symbol for the brain mut~tnt listed here (mbd) stands for m u s h r o o m - b o d i e s
d e r a n g e d
. The results on the courtship tests are summarized here (by the symbolic entries in the matrix) and are compared
to the effects of these genetic yariants on some of the more general aspects of conditioning in D. m e l a n o g a s t e r (for more details see Tully,
this issue). A + entry means that the learning or memory tested was normal or nearly so. A means that the same test revealed demonstrably
worse than normal (wild-type) performance though not necessarily a totally absent response. ( - ) means that aberrant memory is inferred
(but has not been directly tested), because learning per se is abnormal. In the tests of general associative conditioning using electric shocks
and odors (again, refer to Tully's accompanying article), an absence of or defect in "learning" (i.e., a - entry) often seems to mean that
the effects of training tend to decay so rapidly that no learning is revealed; that is, several of these mutants are unable to learn but are
extremely fast forgetters
(for discussion see Tempel et al., 1983; Dudai, 1983)
. The same genetic variants may cause a given component of
courtship "learning" (as formally distinct from " m e m o r y " in terms of the column headings in this table) to be subnormal for the same
reason. NT, not tested.
b rut males apparently cannot discriminate between mated females and virgin females in the first place, in that they court both kinds of females
with approximately the same vigor, unlike what is found for wild-type males or those expressing the other conditioning mutations listed here
(Gailey et al., 1984). Therefore the defective aftereffect exhibited by rut males may not be a learning abnormality.
a m n flies do show subnormal "habituation indices," but they are also lower than normal in their " n a t i v e " responsiveness to sugar stimuli.
Thus, their learning defect in this particular case is difficult to assess unambiguously
(Duerr and Quinn, 1982)
d Ddet~ mutants are defective in associative learning per se. The absence of an influence of this gene on memory is inferred from testing the
effects of D d c variants in heterozygous condition on associative conditioning (Tempel et al., submitted for publication).
" Since D d c t~ mutants have been shown to have much higher than normal thresholds for their basic responsiveness to sugar stimuli, it would
be difficult to test for aberrant habituation or sensitization using such stimuli (Tempel et al., 1984).
Something that should be borne in mind concerning these outlandish
ideas is that the brain of D. melanogaster is not quite so completed in its
development as one might think. The mushroom bodies in the dorsal brain
of young adults initiate a new round of development beginning at eclosion.
The fiber number in the peduncles of these bodies increases by at least
15% during the first week of adult life
. This increase
depends on environmental inputs, including stimuli received by a fly from
has also demonstrated that, whereas
the absence of visual inputs does not block this increase in fiber number,
disturbing antennal input (the entry point for song stimulation) does.
Therefore we must for the moment, at least, entertain the unlikely
possibility that a young male's aphrodisiacs elicit auditory stimuli which
may modulate postpupal brain development and which, in turn, may
influence his later production of courtship song. We already know (C. P.
Kyriacou and J. C. Hall, unpublished) that males kept in isolation from
egg to adult produce superficially normal pulse and hum song. We have
not yet tested whether this song is normal in all its various components.
Our final phenomenon which involves memory in courtship behavior
is the priming effect of artificial songs on females
Kyriacou and Hall, 1984)
. Here there is an increased reproductive success
associated with the previous experience of females. In the laboratory, 2
rain of intensive prestimulation of females with courtship songs leads to
a priming effect where female receptivity is enhanced for approximately
1-3 min. In the real world, courtship interactions may be brief, and it is
not difficult to imagine that the female's receptivity to such stimuli may
linger a little after initial reception. The benefits of this are obvious for
the courting male, as this will increase the probability that he will mate.
For the female, the adaptiveness of such priming and "summation" of
may be related to her " c o y n e s s . " The female's
only clue as to the fitness of the male is his courtship behavior, and thus
her initial reluctance to mate allows her to screen the male. Consequently
any kind of simple, even transient memory may help her to integrate the
relevant information from an elaborate display. Here, we are back to the
perhaps more conventional view of the informational content of courtship
signals as opposed to their manipulative potentials.
It is worth pointing out that an early study by Schilcher (1976b)
suggested that only the hum component of the song, not an invariant pulse
song, was an effective primer for the female. Thus hums appeared to
" a r o u s e " the female, whereas the pulse could be considered as the
species-specific "trigger" for mating.
Kyriacou and Hall (1984)
confirmed Schilcher's findings but noted that when a more natural
speciesspecific pulse song, including the rhythmic component, was applied to
females, this p r o v e d to be an e v e n m o r e effective priming stimulus than
h u m song in enhancing subsequent mating success. T h e r e f o r e both pulses
and h u m s can be r e m e m b e r e d by the female. Finally. in passing, it should
be reiterated that a species-specific, cycling pulse song also gives faster
mating speeds c o m p a r e d to an invariant pulse song, or one with an
inappropriate r h y t h m , w h e n these sounds are used to stimulate males and
( K y r i a c o u and Hall, 1982)
One m o r e implication o f the m a n y speculative ideas discussed h e r e
is that these kinds o f e x p e r i m e n t s n e e d to be e x t e n d e d b e y o n d the
labo r a t o r y into field studies or at least m o r e natural settings. This will permit
a t r u e r a s s e s s m e n t o f any functional significance or adaptive value which
can be inferred f r o m the laboratory. Of course, it could be that some of
these observations are rather d r e a r y l a b o r a t o r y artifacts. N e v e r t h e l e s s ,
r e p r o d u c t i o n - - w h e t h e r or not it is a n a l y z e d in an artificial s e t t i n g - - i s
implicitly tied to fitness. T h e mutationally induced abnormalities which
disturb the effects o f previous e x p e r i e n c e on courtship b e h a v i o r bring
these genes into focus as having e v o l v e d to do m o r e than m e r e l y aid flies
to avoid electric shocks. Table I presents a s u m m a r y o f these genetically
c a u s e d abnormalities in artificial elements of conditioned b e h a v i o r and
in t h o s e which we d e e m to be quasirealistic.
L e a r n i n g in D r o s o p h i l a would a p p e a r to h a v e b e e n f a v o r e d by natural
(Hewitt et aI., 1983)
, and t h e r e f o r e it is important to identify
those p r o c e s s e s in the life history of the fly which are particularly sensitive
to behavioral plasticity. We would suggest that r e p r o d u c t i v e b e h a v i o r is
one o f those p r o c e s s e s ; " r e w a r d learning" involving nutritive
(Tempel et al., 1983)
is likely to be another one. F u r t h e r
studies which intensify the analyses o f these behaviors, their neural
underpinnings, and the genes themselves should continue to p r o d u c t
valuable information as t h e y are carried out in parallel.
Aceves-Pifia , E. O. , Booker , R. , Duerr , J. S. , Livingstone , M. S. , Quinn , W. G. , Smith , R. F. , Sziber , P. P. , Tempel , B. L. , and Tully , T. P. ( 1983 ). Learning and memory in Drosophila studied with mutants . Cold Spring Harbor Syrup . 48 : 831 - 840 .
Arnold , A. P. ( 1980 ). Sexual differences in the brain . Am. Sci . 68 : 165 - 173 .
Baker , B. S. , and Belote , J. M. ( 1983 ). Sex determination and dosage compensation in Drosophila melanogaster . Annu. Rev. Genet . 17 : 345 - 393 .
Bastock , M. ( 1967 ). Courtship, A Zoological Study , Heinemann, London.
Bastock , M. , and Manning , A. ( 1955 ). The courtship of Drosophila melanogaster . Behaviour 8 : 85 - 111 .
Booker , R. , and Quinn . W. G. ( 1981 ). Conditioning of leg position in normal and mutant Drosophila . Proc. Natl. Acad. Sci. USA 76 : 3940 - 3944 ,
Caryl , P. G. ( 1979 ). Communicationby agonistic displays: What can game theory contribute to ethology? Behaviour 68 : 136 - 139 .
Caryl , P. G. ( 1982 ). Animal signms: A reply to Hinde . Atom. 3ehav . 30 : 240 - 244 .
Connolly , K. , and Cook , R. ( 1973 ). R~ectio~n behaviours by female Drosophila melanogaster: Their ontogeny, causality and effects upon the behaviour of the courting male . Behaviour 52 : 155 - 171 .
Cook , R. ( 1977 ). Behavioral role of the sex combs in Drosophila melanogaster and Drosophila simulans . Behav. Genet . 7 ; 349 - 358 .
Cook , R. , and Cook , A. ( 1975 ). The attractiveness to males of female Drosophila melanogaster: Effects of mating, age, and diet . Anim. Behav . 23 : 521 - 526 .
Davis , R. L. , and Kauvar , L. M. ( 1984 ). Drosophila cyclic nucleotide phosphodiesterases . Adv. Cyclic Nucleotide Res . 16 : 393 - 402 .
Dawkins , R. , and Krebs , J. R. ( 1978 ). Animal signals: Informataon or manipulation? In Krebs , J. R. , and Davies , N. B. (eds.), Behavioral Ecology , Sinauer Associates, Sunderland, Mass., pp. 282 - 309 .
Delcomyn , F. ( 1980 ). Neural basis of rhythmic behavior in animals . Science 210 " 492 - 498 .
Dudai , Y. ( 1983 ). Mutations affect the storage and usage of memory differentially in Drosophila . Proc. Natl. Acad. Sci. USA 80 : 5445 - 5448 .
Duerr , J. S. , and Quinn , W. G. ( 1982 ). Three Drosophila mutations that block associative learning also affect habituation and sensitization . Proc. Natl. Acad. Sci. USA 79 " 3646 - 3650 .
Ewing , A. W. ( 1977 ). Communication in Diptera . In Sebeok, T. A. (ed.), How Animals Communicate , Indiana University Press, Bloomington and London, pp. 403 - 417 .
Ewing , A. W. ( 1983 ). Functional aspects of Drosophila courtship . Biol. Rev . 58 : 275 - 292 .
Galley , D. A. , Jackson , F. R. , and Siegel , R. W. ( 1982 ). Male courtship in Drosophila: The conditioned response to immature males and its genetic control . Genetics 102 : 771 - 782 .
Galley , D. A. , Jackson , F. R. , and Siegel , R. W. ( 1984 ). Conditioning mutations in Drosophila melanogaster affect an experience-dependent behavioral modification in courting males . Genetics 106 : 613 - 623 .
Hall , J. C. ( 1979 ). Control of male reproductive behavior by the central nervous system of Drosophila: Dissection of a courtship pathway by genetic mosaics . Genetics 92 : 437 - 457 .
Hall , J. C. ( 1982 ). Genetics of the nervous system in Drosophila . Q. Rev. Biophys . 15 : 223 - 479 .
Hall , J. C. , Gelbart , W. M. , and Kankel , D. R. ( 1976 ). Mosaic systems . In Ashburner, M., and Novitski , E. (eds.), The Genetics and Biology of Drosophila , Vol. la, Academic Press, London, pp. 265 - 314 .
Hall , J. C. , Siegel , R. W. , Tompkins , L. , and Kyriacou , C. P. ( 1980 ). Neurogenetics of courtship in Drosophila . Stadler Genet. Symp . 12 : 43 - 82 .
Heisenberg , M. ( 1980 ). Mutants of brain structure and function: What is the significance of the mushroom bodies for behavior? In Siddiqi , O. , Babu , P. , Hall , L. M. , and Hall , J. C. (eds.), Development and Neurobiology of Drosophila , Plenum Press, New York, pp. 373 - 390 .
Hewitt , J. K. , Fulker , D. W. , and Hewitt , C. A. ( 1983 ). Genetic architecture of olfactory discriminative avoidance conditioning in Drosophila melanogaster . J. Comp. Psychol . 97 : 52 - 58 .
Hinde , R. ( 1981 ). Animal signals: Ethological and game-theory approaches are not incompatible . Anim. Behav . 29 : 535 - 542 .
Hotta , Y. , and Benzer , S. ( 1976 ). Courtship in Drosophila mosaics: Sex-specific foci for sequential action patterns . Proc. Natl. Acad. Sci. USA 73 : 4154 - 4158 .
Jackson , F. R. , Gailey , D. A. , and Siegel , R. W. ( 1983 ). Biological rhythm mutations effect an experience-dependent modification of male courtship in Drosophila melanogaster . J. Comp. Physiol . 151 : 545 - 552 .
Jallon , J. -M. , and Hotta , Y. ( 1979 ). Genetic and behavioral studies of female sex appeal in Drosophila . Behav. Genet . 9 : 257 - 275 .
Konishi , M. , and Gurney , M. E. ( 1982 ). Sexual differentiation of brain and behavior . Trends Neurosci. 5-20-23.
Konopka , R. J. ( 1984 ). Neurogenetics of Drosophila circadian rhythms . In Huettel, M, D. (ed,), Evolutionary Genetics of lnvertebrate Behavior , Plenum Press, New York (in press).
Konopka , R. J. , and Benzer , S. ( 1971 ). Clock mutants of Drosophila melanogaster . Proc. Natl. Acad. Sci. USA 68 : 2112 - 2116 .
Konopka , R.. Wells , S. , and Lee , T. ( 1983 ). Mosaic analysis of a Drosophila clock mutant . Mol. Gen. Genet . 190 : 284 - 288 .
Kyriacou , C. P. ( 1981 ). The relationship between locomotor activity and sexual behaviour in ebony strains of Drosophila melanogaster . Anim. Behav . 29 : 462 - 471 .
Kyriacou , C. P. , and Hall , J. C. ( 1980 ). Circadian rhythm mutations in Drosophila melanogaster affect short-term fluctuations in the male's courtship song . Proc. Natl. Acad. Sci. USA 77 : 6729 - 6733 .
Kyriacou , C. P. . and Hall , J. C. ( 1982 ). The function of courtship song rhythms in Drosophila . Anim. Behav . 3 : 794 - 801 .
Kyriacou , C. P. , and Hall , J. C. ( 1984 ). Learning and memory mutations impair acoustic priming of mating behaviour in Drosophila . Nature 308 : 62 - 65 .
Livingstone , M. S , and Tempel , B. L. ( 1983 ). Genetic dissection of monoamine synthesis in Drosophila . Nature 303 : 67 - 70 .
Manning , A. ( 1967 ). The control of sexual receptivity in female Drosophila . Anim. Behav . 15 : 239 - 250 .
Markow , T. A. , and Hanson , S. J. ( 1981 ). Multivariate analysis of Drosophila courtship . Proc. Natl. Acad. Sci. USA 78 : 430 - 434 .
MacLusky , N. J. , and Naftolin , F. ( 1981 ). Sexual differentiation of the central nervous system . Science 211 : 1294 - 1303 .
Maynard Smith , J. ( 1956 ). Fertility, mating behaviour and sexual selection in Drosophila subobscura . J. Genet . 54 : 261 - 279 .
Menzel , R.. and Erber , J. ( 1978 ). Learning and memory in bees . Sci. Am . 239 :(1): 102 - 110 .
Quinn , W. G. , and Greenspan , R. J. ( 1984 ). Learning and courtship in Drosophila: Two stories with mutants . Annu. Rev. Neurosci . 7 : 67 - 93 .
Quinn , W. G. , Sziber , P. P. , and Booker , R. ( 1979 ). The Drosophila memory mutant amnesiac . Nature 277 : 212 - 214 .
Schilcher , F. von (1976a). The role of auditory stimuli in the courtship of Drosophila melanogaster . Anim. Behav . 24 : 18 - 26 .
Schilcher , F. yon (1976b). The function of pulse song and sine song in the courtship of Drosophila melanogaster . Anim. Behav . 24 : 622 - 625 .
Schilcher , F. von, and Hall, J. C. ( 1979 ). Neural topography of courtship song in sex mosaics of Drosophila melanogaster . J. Comp. Physiol. A 129 : 85 - 95 .
Shotwell , S. L. ( 1983 ). Cyclic adenosine 3' :5'-monophosphate phosphodiesterase and its role in learning in Drosophila . J. Neurosci , 3 : 739 - 747 .
Siegel , R. W. , and Hall , J. C. ( 1979 ). Conditioned responses in courtship of normal and mutant Drosophila . Proc. Natl. Acad. Sci. USA 76 : 3430 - 3434 ,
Spieth , J. T. , and Ringo , J. M. ( 1983 ), Mating behavior and sexual isolation in Drosophila . In Ashburner, M. Carlson, J. L. , and Thompson , J. N , (eds.), The Genetics andBiology of Drosophila , Vol. 3c , Academic Press, London, pp. 223 - 284 .
Szabad , J. , and Fajszi , C. ( 1982 ). Control of female reproduction in Drosophila: Genetic dissection using gynandromorphs . Genetics 100 : 61 - 78 .
Technau , G. ( 1984 ). Fiber number in the mushroom bodies of adult Drosophila melanogaster depends on age, sex, and experience . J. Neurogenet . 1 : 113 - 126 .
Tempel , B. L. , Bonini , N. , Dawson , D. R. , and Quinn , W. G. ( 1983 ). Reward learning in normal and mutant Drosophila . Proc. Natl. Acad. Sci. USA 80 : 1482 - 1486 .
Tempel , B. L. , Livingstone , M. S. , and Quinn , W. G. ( 1984 ). Mutations in the dopadecarboxylase gene affect learning in Drosophila . Proc. Natl. Acad. Sci. USA 81 : 3577 - 3581 .
Tompkins , L. , and Hall , J. C. ( 1981 ). The different effects on courtship of volatile compounds from mated and virgin Drosophila females . J. Insect Physiol . 27 : 17 - 21 .
Tompkins , L. , and Hall , J. C. ( 1983 ). Identification of brain sites controlling female receptivity in mosaics of Drosophila melanogaster . Genetics 103 : 179 - 195 .
Tompkins , L. , Gross , A. C. , Hall , J. C. , Gailey , D. A. , and Siegel , R. W. ( 1982 ). The role of female movement in the sexual behavior of Drosophila melanogaster . Behav. Genet . 12 : 295 - 307 .
Tompkins , L. , Siegel , R. W. , Gailey , D. A. , and Hall , J. C. ( 1983 ). Conditioned courtship in Drosophila and its mediation by chemical cues . Behav. Genet . 13 : 565 - 578 .
Wright , T. R. F. ( 1977 ). The genetics of dopa decarboxylase and c~-methyl dopa sensitivity in Drosophila melanogaster . Am. Zool . 17 : 707 - 721 .