Analysis of Heterogeneity in Complex Characters 363 



effects of quantitative accentuation of direction of 



VARIATION IN ELEMENTS OF PATTERN. 



In this application of the idea of modification through quantitative accumula- 

 tion or subtraction I shall deal with changes involving only the directions of 

 variation in simplest characters, the present indivisible elements of the pronotal 

 pattern. It has been shown in an earlier section how the elements of the 

 pronotal pattern " vary " in diflFerent directions, and how this results in the pro- 

 duction of quite diverse combinations or patterns. It has further been shown 

 that the biotype groups are mainly dependent upon this difference in combina- 

 tion of the elements into stable, separated groups, capable of isolation by 

 modern methods. In this section is considered the possibility of modifying 

 quantitatively the fluctuations in the lines of variation present in the simplest 

 characters. 



The problem can be attacked by attempting to change biotype 4 into biotype 5, 

 by the attempt to change the direction of fusions between (a) and (d) from one 

 in an antero-mediad direction to the anterior end of (a), to a strictly transverse 

 direction as in biotype 5, and at the same time obtain by selective reduction a 

 loss of the anterior and posterior marginal spots of biotype 4. In these experi- 

 ments only homozygous-acting, carefully pedigreed stocks, reared under uniform 

 conditions, can be used, and any deviation introduces errors that obscure or 

 falsify the results. 



Many attempts have been made to accomplish the result proposed by quanti- 

 tative change. Thus far no success has rewarded these attempts. At the outset 

 difficulty is encountered in finding the proper fluctuations upon which to work. 

 In many series I have repeatedly failed because the supposed fluctuation in the 

 desired direction proved to be only a somatic disturbance and not at all a gametic 

 change, while others proved to be gametic in character, but of too low capacity 

 for use in accentuation. So, likewise, the attempt to remove the anterior and 

 posterior fusion lines did not succeed, although it was possible to maintain with 

 ease much reduced states of these. In no case did these prove to be wanting in 

 the selected strains. Crosses with these and some other biotype always showed 

 exactly the behavior of a biotype 4 when crossed with the test strains. 



The same result has thus far attended my efforts to transmute biotype 7 into 

 either 9, 10, 11, or 12, by quantitative accentuation. The best efforts of which 

 I am capable have been put upon tests of this kind, and I have everywhere been 

 rewarded by complete failure to attain by quantitative change any transmutation 

 from one biotype to another. An example of the difficulty attending the work- 

 ing out of experiments of this type is found in the experiments wherein I was 

 trying to change biotype 7 into either 1, 10, 11, or 12, and tests were made of the 

 real nature of the accentuation shown by crossing back upon another strain of 

 type 7. In 11 instances it happened that there appeared in such tests biotypes 

 of the desired kind, but in all it was found that the test strain had carried in 

 " unseen condition " the pronotal form-factor for the "' muUilineata form," and 

 this crossed with type 7 always by synthetic combination gives fixed pure- 

 breeding stocks in F, of biotypes 9, 10, and 11, rarely of 12. The changes 

 found in these tests were not at all a product of quantitative action, but a known 

 result following the presence of a definite factor when introduced into the game- 

 tic complex. To those familiar by actual experience with the delicate nature of 

 some of the gametic factorial effects, the above will appear most obvious. It is 



