The speed r.t vhich this coverinjj took place Virns rennrkcble. 

 In one or t:vo days a. fur visible to the naked eye would form. 



Upon moultinf^ the larvae Y;ould of course be suddenly free of 

 this grovrth. A speedy succession of moultings vras therefore of 

 great advantage anc. t'ne 1919 promptness of moulting left much to 

 be desired. 



In 1920 no difficulty \rc.s had in this respect probably because 

 wc nanrged to keep the wrter in the tanlcs inuch cleaner. By using 

 •the canvas covering from the beginning to the end diatoms also did 

 not give any noticeable trouble. The cleanliness of the v/ater is 

 chiefly determined by the quantity (perhaps also somewhat by the 

 type) of food jjjivon and by the rate, of water change,. 



As regards the needed quantity of food I did not have the 

 least information and by lack of data on this score "one is apt to 

 give too much for fear that the larvae vxill suffer and not grov; 

 fast enough. Establishing the right amount is the result of 

 experience. Too little is of course wrong, too much will soon bring 

 on -a dirtying of the v/ater and the forming of thread like, yellow- 

 ish masses which vrould grow to enormous proportions in remarkably 

 short periods. These consist .-^f the threads of fungi, diatoms and 

 bact^eria charged v/ith i?ifusoria and food particles o 



There being but little knovm of the food which larvae normally 

 take in nature, feeding \inder artificial conditions has been exclu- 

 sively with animal-food such as meat, liver, fish, even hard 

 boiled egg, Jelly-fish and meat of crustaceans, all in finely 

 .ground condition. It has been impossible to determne hovrever 

 whether a preference existed for any of these or v;hethor a notice- 

 ably faster gr-T.-t?! resulted. 



During my experiments I generally used finely groxon-'' crab 

 (Carcinus moenas) these v;ore in their entirety ground in a ricat 

 grinder to a mushlike consistency, placed in a -strainer and sifted 

 out into the water so that only the finer particles would leave and 

 the courser ones, mostly reri^ins of shells, stay in the strr.iner. 



Every tvro hours, a quantity/ '"^f about 25 cubic cm. -f this pap 

 v/ould be released in each small tani:, later these quantities v/ere 

 reduced. One must realize that c. ,,roat part of this food consisted 

 of 'shell and that a certain proportion would soon exit from the 

 tanlc tl'O'ou^h the screens of which the mesh v/as much wider than that 

 of the strainer. Ground fish, as for instance young herring, also 

 gave good results. As the larvae grew, the food vras generally 

 changed to a more so-lid variety as ground oysters or ground fish 

 of more solid meat. Another measure which .x.terially aided in 

 changing the vrater was the introduction of the larger v.'indovrs and 

 the opportunity to clean these -^ftener» 



140 



