Localization
Lobster 5HT neurons originally were localized within
the ventral nerve cord of the lobster central nervous
system, (CNS) using immunocytochemical methods
(Beltz and Kravitz 1983). The results showed that there
were approximately 120 5HT neurons, widely distrib-
uted among the ganglia of the nerve cord. Recently,
additional subsets of neurons have been found that
show immunostaining for 5HT, but only after incuba-
tion of tissues in low concentrations of the amine. It is
not certain, however, whether these neurons also con-
tain the synthetic enzyme, tryptophan hydroxylase
(Beltz et al. 1998; Musolf and Edwards 1999). Such
neurons, which likely contain high levels of a 5HT
transporter, are capable of release of the amine (Beltz
et al. 1998; Musolf and Edwards 1999). Therefore, de-
spite the possible absence of the key synthetic enzyme,
these neurons may be utilizing 5HT as a neurotrans-
mitter or neurohormone. It is not clear how such cells
should be categorized. For the original endogenous
5HT-immunostaining cells, however, it was easy to
demonstrate that they were organized into distinct
subgroups on morphological grounds. The groups in-
cluded short and long-distance projecting interneurons,
local projection interneurons, and neurosecretory neu-
rons. The most detailed physiological studies have been
carried out with the neurosecretory neurons, but at least
one key supraesophageal (brain) interneuron, the de-
utocerebral giant cell also has received considerable at-
tention, at least in cray®sh (Sandeman and Sandeman
1987, 1994, 1995).
Identi®cation of amine neurons
Strong evidence that a neuron is serotonergic involves
®rst developing a method to routinely ®nd the same cell
in dierent animals, and then to show by biochemical
techniques that the cell actually contains the amine, and
if possible, also contains the synthetic enzyme tryptop-
han hydroxylase. In lobsters, this has been done for the
neurosecretory subset of serotonergic neurons, which
are the focus of this review (Beltz and Kravitz 1987). The
technique used was ®rst to carry out immunocyto-
chemical studies to show the approximate location
within ganglia of neuronal somata staining for amines
(Beltz and Kravitz 1983). With the antibodies used, the
arbors of neurons also stained well, allowing tracing of
axons into nerve bundles and tracts within the ventral
cord. With the immunocytochemical studies as a guide,
neurons were identi®ed by stimulating nerve bundles
containing the axons of the cells in question, while re-
cording with intracellular electrodes from cell somata in
the vicinity of the candidate immunostained cell. Cell
bodies that showed action potentials when activated in
this way were ®lled with electron dense or ¯uorescent
dye markers through the intracellular electrode. There-
after, tissues were processed to allow visualization both
of the injected dye and the amine-associated immuno-
reactivity. Once candidate neurons were routinely iden-
ti®able in this way, their physiological properties were
explored. Such studies ultimately allowed identi®cation
of the neurons using physiological criteria alone (see
below). To con®rm that the immunostained cells actu-
ally contained 5HT, the somata were microdissected
from nerve cords to measure their amine and amine
biosynthetic enzyme contents, and thereby to con®rm
their identity. Using this method we have shown that the
average 5HT concentration in neurosecretory neuron
cell bodies is 0.4±0.6 mmol l
)1
(Siwicki et al. 1987).
5HT-neurosecretory neurons also contain the peptide
proctolin at approximately 20 lmol l
)1
(Siwicki et al.
1987), but the amine and peptide phenotypes of these
cells ®rst appear at widely dierent times in development
(Beltz and Kravitz 1987).
Morphological and physiological properties
of 5HT-containing neurosecretory neurons
Morphology
The above studies identi®ed two pairs of large 5HT-
containing neurosecretory neurons: one pair found in
the 5th thoracic (T5-5HT cells) and a second in the 1st
abdominal (A1-5HT cells) ganglia. The T5- and A1-5HT
cells are important in postural and escape neuronal
circuitries (see below), and serve as the principal and
possibly the only source of 5HT reaching the circulation.
Peripheral targets of amines, like muscles (Glusman and
Kravitz 1982; Dixon and Atwood 1989; Goy and Kra-
vitz 1989) and sensory neurons (Pasztor and Bush 1989),
have no serotonergic innervation, and are supplied ex-
clusively via the circulation. Once a reliable method
existed for ®nding these neurons, intracellular injections
of horseradish peroxidase, lucifer yellow, or cobalt
chloride allowed mapping of their morphology (Beltz
and Kravitz 1987).
The serotonin-containing neurosecretory neurons are
suprasegmentally organized and both pairs show similar
morphological features. The general plan is as follows. A
single process emerges from the cell body, which gives
o an elaborate arbor of branches and endings in the
neuropil region of the hemiganglion containing the so-
ma. One branch, the ``axon'', travels in a rostral direc-
tion in an anterior lateral nerve bundle for one segment
(to the next ganglion), turns medially, and then ascends
the nerve cord through all rostral thoracic ganglia to
ultimately end in the subesophageal ganglion. In the
neuropil of each anterior ganglion, this branch gives o
a main process that divides: one branch leaves the neu-
ropil via the 2nd nerve root to terminate in one or more
extensive ®elds of neurosecretory endings along the root;
the second branch yields an arbor of endings in the
neuropil of the ganglion (Fig. 1). Amines reach periph-
eral targets like exoskeletal muscles and sensory neurons
by release into the hemolymph from the endings along
224