1983; Friesen 1989), feeding (cf. Lent 1985; Schachtner
and BraÈunig 1993; Yeoman et al. 1994), and ®ghting
(Adamo et al. 1995; Edwards and Kravitz 1997)
behaviors. At the start we must acknowledge that it is a
formidable challenge to construct a model of how
amine neurons function in aggression. This is true
despite the growing numbers of experimental results we
and others are obtaining: in behavioral studies with
living lobsters and cray®sh; in physiological experi-
ments de®ning the intrinsic properties of amine
neurons; in systems studies exploring the function of
these neurons in circuits; and in combined behavioral/
physiological studies ®nding changes in the properties
of amine neurons or their targets that relate speci®cally
to changes in social status.
Watching amine neurons at work during ®ghting
behavior would be an important way to learn more
about their role in aggression. Unfortunately, so far no
recordings have been made from amine neurons in
awake behaving lobsters, as have been done in verte-
brate (Veasey et al. 1995, 1997; Leung and Mason 1999)
and in a few invertebrate (Kupfermann and Weiss 1982;
Schachtner and BraÈunig 1993; Yeoman et al. 1994)
systems. In the absence of such recordings, the specu-
lations we oer rest on what we now know from in vitro
studies about amine neurons in crustaceans, and on
information and suggestions gathered from studies of
other invertebrate and vertebrate neuron systems.
The model proposed here borrows conceptually from
the hypothesis originally put forth by Phoenix et al.
(1959) that steroid hormones serve two distinct roles in
modulating behavior: one organizational and the other
activational. Important modi®cations of the original
hypothesis were made by Arnold and Breedlove (1985)
who pointed out that the divisions between these two
roles were not sharp. The hypothesis was based on
studies of testosterone action on mating behavior in
newborn and adult guinea pigs. Phoenix et al. (1959)
con®rmed and extended studies showing that gonadal
steroids had actions at around the time of birth, that
were essential for the much later in development, be-
havioral responses evoked by release of the same ste-
roids. It was suggested that the initial exposure had an
organizational eect, a carving out of future gonadal
steroid responsive territories within the brain and other
body tissues. Much later in development, upon release of
the gonadal steroids again, this time in sexually mature
animals, the appropriate male and female behavioral
patterns are triggered. The later responses represent the
activational component of steroid hormone action.
The organizational role of amines
Amines like 5HT also are found early in development of
the nervous system in many species of animals (for re-
view see Lauder 1990; also see Wallace and Lauder 1983;
Wallace 1985; Glover et al. 1987; Aitkin and Tork 1988;
Konig et al. 1988; Goldberg and Kater 1989; Beltz et al.
1990, 1992). In lobsters, for example, the ®rst visible
5HT immunostaining is seen at about 10% of embryonic
life, and the complete set of 5HT neurons is found by
50% of development (Beltz et al. 1992). These times are
well before most of the targets of the 5HT neurons have
formed. Why are amines seen so early in development?
A confusing, but compelling, literature suggests that
amines and other classical neurotransmitters have im-
portant roles in development that far precede their later
roles as neurotransmitters and neurohormones (for re-
views see Lauder 1990; Buznikov et al. 1996). These
earlier roles involve the earliest cleavage divisions of the
embryo (or even gamete formation), and include mo-
rphogenetic roles in cell movement and cell shape
changes during early embryogenesis and later roles when
the nervous system begins to form and dierentiate (for
review see Buznikov et al. 1996). Various non-neuronal
sites, such as yolk granules and notochord, appear ca-
pable of synthesis of 5HT in the early embryo, at least in
some species (see Buznikov et al. 1996). Highly speci®c
serotonin transporters and particular receptor subtypes
show transient patterns of expression in the early em-
bryo and early nervous system as well (Bennett-Clarke
et al. 1993, 1996; Lebrand et al. 1998). The ®rst neurons
expressing amine transmitters appear well before their
targets are formed (Lauder et al. 1982; Wallace and
Lauder 1983; Wallace 1985). Thus, serotonergic neu-
rons, expressing the transmitter phenotype, are seen
growing through primitive epithelial layers that have not
yet dierentiated to form neurons. Indeed, 5HT, possi-
bly released from growth cones, has been suggested to
serve critical roles in the growth and dierentiation of
certain of its target neurons and in the activation of glial
cells that then secrete growth and dierentiation factors
of their own (Haydon et al. 1987; Goldberg and Kater
1989; Lauder 1990). Too little, or too much 5HT, both
lead to morphogenetic abnormalities (Goldberg and
Kater 1989; Cases et al. 1996; Upton et al. 1999). For
example, in mice missing the monoamine oxidase A
gene, which results in elevated levels of 5HT and nor-
epinephrine in the brain, various abnormalities are seen.
The cortical barrel ®elds fail to develop (Cases et al.
1996) and retinal ganglion cell ®bers that ordinarily
segregate into layers in the lateral geniculate body, fail to
do so in the mutant animals (Upton et al. 1999). In-
hibiting the synthesis of 5HT at early developmental
stages prevents the formation of these abnormalities
(Upton et al. 1999). In monoamine oxidase A knock-out
mutants, 5HT immunostaining is seen in a subpopula-
tion of ganglion cell neurons that contain a serotonin
transporter, while in control animals immunostaining is
not seen (although the transporter is expressed in the
same transient manner as in the mutant). The machinery
for loading the amine into vesicles also is present in some
of these neurons, as is a protein thought to be involved
in vesicular binding of 5HT (Upton et al. 1999). Thus,
the cells that take up the amine may be capable of re-
leasing it as a borrowed transmitter in inappropriate
places or amounts in the mutant animals (Cases et al.
231