change during ®ghts, under the in¯uence of synaptic
inputs of the types we have identi®ed. Changes in the
rates of ®ring of serotonergic neurons in behaving ani-
mals have been recorded in other species (Kupfermann
and Weiss 1982; Schachtner and BraÈunig 1993; Yeoman
et al. 1994). Invariably these changes relate speci®cally
to the behaviors the neurons originally were postulated
to be involved with. In some cases, the changes in ®ring
anticipate the behavior. For example, in locusts, ser-
otonergic neurons of the satellite nervous system begin
®ring tens of seconds before the initiation of feeding
behavior in the animals (Schachtner and BraÈunig 1993).
With the information gathered so far from in vitro
studies, we can propose dierent scenarios for how
®ghting behavior might alter amine neuron function,
and in turn, how changes in amine neuron function
might in¯uence the behavior. Early in ®ghts, when ani-
mals stand tall displaying their claws, increases in the
rates of ®ring of the 5HT cells in both combatants could:
(1) through central actions of amines, maintain animals
in an elevated posture and facilitate the maintenance of
the postural displays; and (2) through 5HT released into
the circulation, increase the strength of muscular con-
tractions (through pre- and post-synaptic facilitatory
actions on exoskeletal junctions). There are several ways
to eect such changes. The ®ring rates of A1-5HT cells
would be immediately increased by blocking the spon-
taneous inhibitory input that holds the ®ring of the cells
in check (see above). Rates also would be increased by
®ring ¯exor commands, which make animals stand in
the elevated posture seen when animals approach each
other, and the increased ®ring, in turn, should enhance
the output of the command through central and pe-
ripheral actions of 5HT (the ``gain-setter'' role; Ma et al.
1992 and see above). Later in ®ghts, after a hierarchy
has been established, dominants alone stand tall as they
approach subordinates. At this point, perhaps only
winners show increased ®ring of their serotonergic ne-
urosecretory neurons. Extensor commands trigger an
opposite, submissive-looking stance, and inhibit the ®r-
ing of serotonergic neurosecretory cells. Such commands
may become more important in the repertoire of losing
animals, since these animals maintain a lowered posture
as they continually retreat from dominants. All of these
speculations, derive from the pool of knowledge gained
from our in vitro studies (see Figs. 3, 4). They would be
greatly strengthened by recording from the A1- or T5-
5HT cells in ®ghting animals, which is a high priority for
future studies.
In the temporal domain, changes in the ®ring rates of
amine neurons might in¯uence how long animals are
willing to ®ght (the ``motivational'' role). As noted
above, the main eect we see of altering amine levels in
behaving lobsters is in how long animals are willing to
®ght, not in whether or not they will ®ght. Amine ac-
tions on most subtypes of target receptors are slower in
on- and oset than the actions of classical transmitters.
Amines trigger these slower eects through activation or
inhibition of a wide variety of 2nd-messenger pathways,
including cyclic nucleotides, phosphatidyl inositides,
calcium and other substances (for review see Hen 1992).
Activation of the cascades triggered by these substances
can produce cellular changes whose durations range
from seconds to minutes if covalent modi®cations of
proteins are involved, or to days if transcriptional reg-
ulation is altered. The durations of amine actions,
therefore, are not determined by the continued presence
of the amine, but by how long it takes target cells to
reverse the changes initiated by activation of the 2nd-
messenger systems. Advance ®ring of amine neurons in
anticipation of behavior, through such pathways, could
prime amine responsive circuitries in order to optimize
their subsequent eectiveness in behavior (see above).
Long-term consequences of changes in status
Long-term changes in the behavior of both animals
follow the establishment of a hierarchy. Thus, winners
are more likely, and losers less likely to win their next
®ghts (Scrivener 1971). Moreover, losing animals will
not ®ght with winners for even a week after their initial
meeting, if animals have been separated from each other
(Karavanich and Atema 1998a; Rutishauser et al. 1999).
If not separated after the ®rst ®ght, the social positions
(winner/loser) and accompanying behavioral patterns
are maintained inde®nitely. The maintenance of the
behavioral status quo, with dominant animals regularly
chasing the retreating subordinates, seems to consolidate
whatever changes were initiated in both animals by their
®rst ®ghts. Ultimately, some of these changes are likely
to be in the form of changes in gene expression. Studies
supporting this idea were described above (Yeh at al.
1996, 1997), when it was seen that 5HT-receptor subtype
distribution changes markedly at particular synaptic
sites after changes in social status. These changes are
slow in onset, however, taking some 2 weeks to reach a
maximum eect. While it has not yet been determined
that they are due to changes in amine receptor gene
expression in lateral giant neurons, that seems a likely
possibility with the very slow time-course of the changes.
It is very unlikely, however, that these represent the
entire spectrum of behaviorally modulated changes
taking place.
An attractive feature of the ``gain-setter'' model for
how amine neurosecretory cells function, is that in ad-
dition to the short-term in¯uences on postural circuit-
ries described above, longer-term changes in the
properties of the amine cells, or in the sensitivity of their
targets to amines, could alter the output of the circuits
in longer-term ways, without any changes required in
the hard-wiring of the system. Any of a variety of
changes in the serotonergic neurons or their targets
could eect such changes. These include: (1) modifying
the membrane ion channel composition of amine neu-
rons, or reducing or enhancing the levels of spontaneous
synaptic input to the cells from inhibitory or excitatory
sources, all of which would change the spontaneous
234