692 MUTANT RACES DERIVED FROM OENOTHERA 



Interchromosomal Correlations. 



Almost all the cultures described above produced a large number 

 of mutants besides the main types. The percentages did not essen- 

 tially differ from those of Lamar ckiana, and the same forms were 

 present in almost every single instance, although the size of the 

 experiments was too small for the expectation of complete series. 

 But all of the primaries occurred, namely, lata, scintillans, cana, 

 liquida, pallescens, spathulata and pulla, also both the accessories, 

 albida and oblonga. Nor did semigigas fail, and nanella occurred either 

 as such or as half-mutants (as in the experiment with cana). From 

 this we may conclude that the activation of one of the primaries 

 or accessories in one of the chromosomes does not change the muta- 

 bility of the others. This condition remains the same as it has been 

 found in the parent species, Oe. Lamar ckiana. 



In this respect the trisomic or specific mutations are essentially 

 different from the homogeneous or homozygotic types, previously 

 described (de Vries 1923, c, p. 221). Oe. blandina, Oe. deserens and Oe. 

 decipiens have produced, until now, almost no mutations, even in 

 cultures of a large size. If crossed with one another, the progeny 

 also shows no mutability, either in the first or in the second genera- 

 tion. Over 350 mostly flowering hybrids have been studied for 

 each of the two, embracing the six possible combinations. Here, 

 therefore, is a clear proof of interference, especially when we com- 

 pare this absence of mutability with the mutability shown in the 

 tables of the present article. Of the homogeneous mutations men- 

 tioned, Oe. blandina belongs to the class of velutina, but Oe. decipiens 

 and Oe. deserens to that of laeta. All of them lack both the zygotic 

 lethals which are characteristic for the parent species. The factor- 

 complexes of all three must be assumed to be represented in the 

 central pair of chromosomes, which also carries the lethal factors, 

 as Boedijn and I (1923, 1924 b) have shown. Those complexes are 

 of a recessive nature. From this we see that the operative condition 

 of these recessives in the central chromosome hinders the mutability 

 of the lateral ones in producing its types. In other words, there 

 exists an interchromosomal correlation between the central rod and 

 the lateral ones, which is absent between the lateral rods among 

 themselves. 



The latent condition of the mutability of Oe. Lamarckiana gigas 

 discovered by means of double crossing (de Vries 1924 b) may be 

 due to such a form of correlation, and this would point to the possi- 



