Eeactions and Pkoducts in Interspecific Ceosses 131 



The most common departure from the routine condition in the crossing of 

 these two species is shown in plate 14, in which the results are graphically given 

 when the cross is signaticollis with the ^c*° and undecimlineata Ac^^. The con- 

 ditions of the cross external to the organisms have been on the average as 

 follows : 



Day .. 

 Night 



Temperature 

 (average). 



" F. 

 101 ±4 

 85 ± 3.5 



Relative humidity 

 (average). 



77 ±6 

 97 ± 3 



Wind movement, av., 440 ft. per min. ; range, 425 to 455. 

 Evaporation (av., 52.3 c. c. per 24 hrs.) : 



Day (7 h. 30 m. a. m. to 5 h. 30 m. p. m.). av., 43.1 c. c. 



Night (5 h. 30 m. p. m. to 7 h. 30 m. a. m.), av., 9.2 c. c. 



These conditions presented evaporation-rates in excess over that found in 

 the normal habitat of undecimlineata and about that present in the habitat of 

 signaticollis. It has made no difference that can be detected whether the stocks 

 came direct from nature or whether the rate of action of ^Ic had been produced 

 in the laboratory, the result is the same. 



The products of the cross are in Fj that the larvae divide into two groups most 

 clearly determined in the second larval stage, yellow and white larvae, and this 

 division is present in the third stage, with both having the pattern-system of 

 undecimlineata dominant ; the larva? are either yellow or white without spots. 

 The numbers are highly variable but on the average are 3 white to 2.7 yellow, 

 essentially a 1 : 1 ratio. The white larvae give only adults that are undecim- 

 lineata in type; the yellow larvae give only mid-types. The two classes of 

 :adults are also present in essentially equal numbers. The F^ undecimlineata 

 types when inbred "breed true" without limits, as pure-breeding races, in 

 aspect showing only the presence of the undecimlineata adult and larval char- 

 acters as long as the breeding is continued. Inbreeding of the mid-types 

 derived from the yellow larvae shows in Fj only the normal and entirely typical 

 array produced by the mating of typical F^ heterozygotes between these two 

 species, and many dozens of tests of these were made between 1906 and 1914, so 

 that we need not concern ourselves further with this portion of the series. The 

 undecimlineata types that are derived in F^ are the portion of the experiment 

 that is of interest. 



In superficial respects in general behavior this experiment is much like the 

 reactions shown in plate 9, in which it was seen that the two F^ classes were 

 different conditions of heterozygosis, and it is on the surface probably the 

 same in this cross. There is one conspicuous difference in the apparently 

 extracted type that comes out in F^ in this cross that is not found in the first 

 case, namely the decreased rate of reaction present in this undecimlineata type. 

 Unlike the normal form which has bji Ac value of about 60 to 64 days, or the 

 Fi heterozygote, which in this cross has the Ac value between 60 and 65 days, 

 the Fi undecimlineata types have the Ac value from 70 to 98 days. The actual 

 length of time between generations is 290 to 335 days, or about a generation in 10 



