No. 636] GAMETIC RATIOS IN DROSOPJIILA 



53 



was at its best when it had fermented for about 24 hours. The 

 optimum amount of drained banana was found to be about 25 

 grams per bottle. This was put upon the bottom of the culture 

 bottle and one gram of paper-toweling strips (about 5. X.7 cm.) 

 was matted down on the top. Pint culture bottles gave a greater 

 output per pair than halfpints. The alcohol method was used 

 more successfully than the old method during the fall and winter 

 of 1916. 



In the spring of 1917 considerable work was done in testing 

 out various media containing starch, sugar, peptone and salts. 

 This method gave good results except that trouble with moulds, 

 was greatly increased. 



In the autumn of 1916, Dr. R. W. Glaser told me of certain 

 culture-media experiments that Mr. Baumberger and he were 

 carrying out with banana infusions and agar.* Dr. Glaser later 

 sent me directions for preparing these media and also some pre- 

 pared tubes. My tests of the method showed that the amount of 

 food was inadequate for general use, although sufficient for the 

 small number of flies that they wished. I increased the con- 

 centration of the media by the addition of sugar, banazia flour, 

 etc., but principally by grinding up and using all the pulp of 

 the bananas, instead of using simply the strained juice. A. 

 .comparison of fresh banana with banana that had been fer- 

 mented before incorporation showed that the fresh banana was 

 superior. Likewise fresh banana was superior to banana raised 

 to the boiling point at any stage of preparation. It was found 

 that yeast should not be distributed throughout the media. 

 Experiments showed that it is advisable to have a very light 

 seeding of yeast confined to the surface of the solidified media. 

 Also it is well to keep the yeast from the margin as much as 

 possible, since fermentation at the sides and beneath the cake 

 makes the cake break loose and rise. Tlie amount of agar waa 

 found to be adequate at 1 per cent. 



It was some months before this method was improved so far 

 that it gave better results than those given by the old method or 

 the alcohol method. In the spring of 1917 it was worked out 

 well enough so that it was substituttnl for the old method in 

 my regular work. By iho wiiit.'i- of 1!>17 it had become quita 



