734 PROCEEDINGS OF THE AMERICAN ACADEMY. 



taken into account, would probably lie somewhere between fifteen and 

 twenty days. 



The normal sexual proportion in this species appears to be one very 

 nearly approaching equality. Mr. Thomas Ordway states (in a manu- 

 script report) that there seems to be a slight excess of males, but that 

 perhaps the proportion would be somewhat changed if larger numbers 

 than he worked with were taken into account. One can say, at least, 

 that neither sex predominates in any marked degree. 



It has been our intention in these experiments to keep external condi- 

 tions as nearly uniform and as nearly optimal as possible, but to breed 

 only from brothers and sisters through a series of generations. The 

 breeding was at first carried on in small glass battery-jars about 8 cm. 

 deep and 6 cm. wide, each covered with a square of window-glass ; later, 

 tumblers of similar dimensions were used, and were found more con- 

 venient because of the flaring top. The jars were kept on a table in the 

 laboratory at ordinary room temperature. 



A single pair of flies (brother and sister) was put in each breeding-jar, 

 and was kept supplied with an abundance of food. This consisted dur- 

 ing generations 1-12 of the A series of fermenting grapes, but in subse- 

 quent generations of the A series, as well as throughout all the other 

 series, it consisted exclusively of banana, in which fermentation was 

 started either by the addition of a little yeast or with a little juice taken 

 from a stock-jar of fermenting banana. 



To guard against the possible production of half-brothers and half- 

 sisters through matings of the same female with two or more males, the 

 following method of forming pairs was adopted and continued up to the 

 year 1904-5. 



Pupae of the same parentage were put together, a pair in each of a 

 number of jars. When the imagoes emerged, if in any jar they proved 

 to be of different sexes, they were left together; if of the same sex, the 

 flies were rearranged so as to bring those of opposite sex together. Thus 

 each female was ultimately placed in a jar with a single male and allowed 

 to remain there during its entire subsequent life, which usually lasted 

 three or four weeks. Most pairs under these circumstances produced an 

 abundance of eggs within three to five days, and pupae were found about 

 a week later attached to the sides or cover of the breeding-jar. These 

 were removed and counted from day to day before they had an oppor- 

 tunity to emerge as imagoes, and the parents were left iu the jar to 

 continue breeding as long as possible. 



Pairs which produced no young by the end of two or three weeks were 



