38 
the water in the tanks was rendered quite turbid, as if 
milk had been poured in. A drop examined under the 
microscope was found to be teeming with spermatozoa in 
active movement. Two days afterwards the Roosebeck 
mussels discharged spermatozoa. From that date on- 
wards, although no further discharge of spermatozoa was 
observed, the eggs were always fertilised. With a very 
few exceptions the eggs were discharged during the night. 
Many of the mussels were actually observed in the act 
of shedding their eggs. 
The embryos flow from the female in a slow distinct 
stream. When not disturbed by currents they settle 
down on the mud close to the parent as an obvious pink 
mass. They remain in this position undergoing the early 
stages of development, which last from eight to twelve 
hours. They then rise to the surface as free swimming 
larvee, and are dispersed by the currents. The duration 
of the free swimming stage was not determined, but the 
larvee remained free swimmers for at least four days. 
The minute size of the larve (they are only from z¢3 to 
zo Of an inch in diameter) prevents accurate observations 
being made for one individual. They can only be kept 
in jars where no circulation is going on, consequently the 
surroundings soon become unfavourable to life. Attempts 
to strain them through fine sieves, and so allow the foul 
water to escape, were not a success, as the larve passed 
through our finest sieves. 
Fertilisation of the eggs has generally been thought to 
take place in the water after they were discharged from 
the parent. That has not been our experience at Piel. 
After the obvious discharge of sperms all the eggs sub- 
sequently extruded were found to ‘be fertilised, even 
though no further discharge of spermatozoa could be 
detected. In several instances the eggs were isolated as 
