550 
BULLETIN OF THE BUREAU OF FISHERIES 
during the winter. This pond is connected with the adjacent sound only during 
exceptionally high tides, which occur generally only a few times during a year, and 
seldom near or in the winter months. It seems rather certain, therefore, that the 
gobies are present during cold weather but are not in open water where they can 
be caught with seines. Several species of Fundulus and Cyprinodon varigatus which 
also inhabit the Mullet Pond, are known to imbed themselves in mud during cold 
weather. Therefore, it seems probable that the gobies also enter the muddy bottom 
during the winter. 
No external structural characters by means of which the sexes may be distin- 
guished have been found. In general, the males range larger in size and are darker 
in color. However, specimens intermediate, both with respect to size and color, 
are nearly always present in collections, making a complete separation of the sexes 
from external characters impossible. In gravid fish the anal papilla, although present 
in both sexes, appears to be larger in the female than in the male. 
SPAWNING 
The information about spawning in Gobiosoma was derived mostly from the 
study of large collections of young. It has not been possible, however, as explained 
elsewhere, to separate definitely the young (less than about 10 mm in length) into 
two groups, representing the two local species, bosci and ginsburgi. Therefore, it 
is not known positively, although it is highly probable, that both species are repre- 
sented among the fry. Neither is it known definitely whether much of the infor- 
mation derived from the study of the collections is applicable to one or both forms. 
The data based on the study of the collections of young give no evidence of two 
predominating spawning periods. Therefore, if two species are represented among 
the fry, the spawning seasons probably occur simultaneously or overlap so fully 
that no distinction may be made either from the size of the young nor from their 
abundance. 
Collections of young Gobiosoma were made from 1927 to 1931. The larvae first 
appeared in the tow in May (the earliest date being May 11, 1929), and throughout 
the summer and fall until December (the latest date being Dec. 6, 1929). The larvae 
were common to abundant from June to September each year but most numerous 
during July and August. During October, November, and December only a few 
scattered ones were secured. 
Among the females in the relatively large collections of adults made in the Mullet 
Pond during August (1930), principally small individuals were in spawning condition, 
the larger ones evidently having spawned out. This condition suggests that the 
principal spawning season was past, and is in general agreement with the situation 
suggested by the data based on the collections of young. Although the fry were 
abundant in the towings during August it may be assumed, that many of those taken, 
particularly the larger ones, were hatched during July. Furthermore, in September 
there occurred a pronounced drop in the number of young. The evidence, therefore, 
is that spawning begins early in May or possibly during the latter part of April, that 
it occurs most abundantly during July and ends except for an occasional late spawner, 
in September. 
The larvae of Gobiosoma were taken over a wide variety of conditions and over a 
comparatively large area, as explained under that section of this paper dealing with 
distribution (p. 558). Since the eggs become attached and do not drift as already 
shown, spawning probably takes place over a large portion of this area and over a 
wide variety of conditions. 
