48 INLAND FISHERIES. 



stars were absent, and not overlooked, in the search on June 22. 

 I feel sure, therefore, that in this locality the starfish first began 

 to set, in considerable numbers, on June 28, or within a day or two 

 of that date. 



Free-swimming larval stars. The first attempt to capture the 

 starfish in the free-swimming stage of their existence was made 

 in the evening of June 27, at Kickemuit. The tow-net w^as dragged 

 at the surface at intervals for two hours in the evening. The con- 

 ditions of the weather were unfavorable, as I afterward learned, 

 and for this reason no young starfish were caught. On the follow- 

 ing morning, June 28, a great number of fry were found swimming 

 at the surface of the water. From this date until July 16 they 

 could be captured at any time, unless the conditions were partic- 

 ularly unfavorable. The few larvae caught on July 16, however, 

 were old, and all of them set in the aquarium during the next 

 twelve hours. After this no more larvae covild be obtained, al- 

 though the next few days were exceptionally favorable. 



It is evident, therefore, that in Kickemuit river the season dur- 

 ing which the starfish larviv set in considerable numbers began 

 about June 28 and ended about July 16, a period of a little less 

 than three weeks. Allowing that the larva^ set three weeks after 

 the spawn is laid (which is the period given by Ingersoll), we 

 may conclude that the beginning of the spaAvning season was 

 about June 7, and the end June 28, while the height of the season 

 was the first, and possibly the second, week in June. From the 

 observations of the sexual glands of the adults, it appeared that 

 the greater part of the spawn was extruded during the 4th and 

 the 16tli of June. The facts, therefore, obtained by both of these 

 methods agree very closelJ^ Through the kindness of Dr. Seitaro 

 Goto, of Tokyo, Japan, who worked in Mr. Agassiz's laboratory at 

 Newport in 1895, I have obtained the following interesting data. 

 For one or two days before the IStli of July, the larva^ had been 

 very numerous. On the 15tli they were decreasing in number 

 and after the 20th of July none were found. At Newport in 1895, 



