132 
BULLETIN or TIIE BUREAU OF FISHERIES 
of time, often occurring for periods of more than fom- months. For the greater 
part of the year they remain attached to hydroids and algae on rocks and piles as 
nonsexual individuals. In this form they are not free-swimming and their occasional 
appearance in surface collections is accidental. In the breeding season certain of ' 
these nonsexual individuals develop eggs in the posterior part of the body (posterior 
to the gizzard), while others develop sperm. Strobilization then occurs, and sexual 
individuals, which immediately become pelagic, are broken off. The females that 
break off carry clusters of eggs in a pouch on the ventral side. The stolons are either 
male or female, the two sexes never developing from the same stalk. Occasionally 
chains of five or six worms, which have not yet separated, may be seen at the surface. 
Free-swimming males and male chains are usually more abundant than the females. 
Alexander Agassiz fully described this alteration of generations in 1862. The sexual 
species of Autolytus are highly phosphorescent and are often extremely numerous 
in the tow. 
Podarke obscura is a very characteristic member of all evening surface collec- 
tions of the summer. On calm, dark nights swarms of them appear at the surface 
in protected harbors. The first specimen taken in 1922 appeared on July 6, the 
last on September 27. In the strong currents about the collecting station the 
occurrence of Podarke was more scattering than is usual, although many were 
carried into the nets during both day and night. In daylight, however, the nvunber 
was much smaller, because at that time the adults seek protection vmder rocks and 
among the Fucus. 
Larvel annelids appear in the plankton at all seasons of the year. During 
the early spring they form almost the only representatives of the benthos in the 
tow. A very small percentage of the species has been worked out, and for that 
reason it has been impossible to identify a large nimiber of the larval forms that 
were taken during the past year. Larval Nereis were very abundant during April 
and May in 1899, 1900, and 1923. These spring forms probably were Nereis 
virens. Another large increase in the latter part of October in each of the years 
recorded may have been N. limbata, although the date is rather late for this species. 
Such conclusions must remain as mere speculation until further data on the breeding 
seasons can be obtained. This can readily be realized if one considers that there 
are six species of the family Nereidse represented at Woods Hole, and larval Nereidi- 
formia have been taken in every month of the year except September. 
Two very characteristic larval annelids appear each year in large numbers. 
The first occurs in late July and continues throughout October. Fewkes has described 
it from Newport as a species of the genus Telepsavus. His identification is doubt- 
ful, however, for no adult of the species has been recorded from this section of the 
Atlantic coast. In 1922 it appeared first on July 26 and continued to be taken until 
October 25. The second larvjB (Lepidonatus squamatus) appeared first on Decem- 
ber 19. Throughout the spring it was taken daily in large numbers. The season 
lasted until the last of April. This fact is rather unusual for Sumner records the 
breeding season as late April, May, and June. An adult female of this species 
taken on February 2, 1923, was filled with ripe eggs. During May and Jime, 1922, 
