846 BULLETIN OF THE BUREAU OF FISHERIES. 



as a cover of the meat and the larvae during the later stages of their growth. In 



Europe several methods have been brought forward which it is claimed will 



secure the desired result. 



Before leaving the subject of fly larvae I beg to call attention to the possi- 



bihty of utilizing for fish food the larvae of other flies, especially those of the 



house fly {Musca domestica) and of the stable flies (of the genera Stomoxys and 



Muscina). Their use would not be attended with the objectionable carrion 



odor, and it is possible that these or some other species might be grown largely 



on vegetable materials. 



SPRATT'S FOODS. 



Several of the Spratt foods have been tried at Craig Brook station, the 

 "fish food" in 1905, the " fibrine fish food " and the "cereal fish food" in 1907. 

 The tests were all made in comparison with chopped hogs' liver. 



In 1905, two lots of brook trout fingerlings of the same origin and character 

 were set apart for the experiment, placed in two ponds which were also of pre- 

 cisely the same character, and kept under the same conditions. Each lot num- 

 bered August I about 20,000. These fish had been fed alike on hogs' plucks 

 and in all respects had been treated alike until the beginning of the test, August 

 5, from which date one lot (no. 1 736) was fed with Spratt's "fish food," while the 

 other (no. 1738), as a control lot, was fed on hogs' plucks, mainly the heart and 

 Ughts. This contrasted feeding, with otherwise identical treatment, was kept 

 up through August 26, having thus continued twenty-two days, after which the 

 feeding on hogs' plucks was resumed. Each morning the ponds were carefully 

 searched, and each dead fish found was at once taken out and recorded. A few 

 days after the test began it was noted that the mortality was increasing in the 

 lot fed on Spratt's food (no. 1 736) , while in the control lot (no. 1 738) it was 

 diminishing. Thus the Spratt's food lot lost during the first ten days of the test 

 as follows: o, o, 3, 4, 5, 6, 11, 11, 13, 10; total, 63; while during the same days 

 the control lot lost 2, 6, 2, o, o, o, 2, i , o, o; total, 13. The disparity in losses con- 

 tinued to increase to the end of the test, and carrying the record forward to the 

 second morning after the close of the feeding we have the following daily losses 

 from August 25 to August 28, inclusive: Of the lot fed on Spratt's food, 38, 69, 

 76, 148; total, 331. Of the control lot, o, o, o, 2; total, 2. The total mortality 

 from the beginning of the test to the second morning after the abandonment of 

 the Spratt's food regimen was, for the Spratt's food lot, 542, and for the control 

 lot, 21. During the next ten days, ending on the morning of September 7, the 

 deaths were: In the Spratt's food lot, 77, 13, 54, 24, 12, 3, 6, 13, 9, 7; total, 

 218; in the control lot there were no losses. By the loth of September the 

 mortaUty in the Spratt's food lot had so far subsided that from that date to the 

 end of the month there were but 9 deaths, against i in the control lot. The 



