LAMARCKIANA SEMIGIGAS. 561 



previously described (7). 0. (blandinax simplex) semigigas(0. perennis) 

 was fertilized, partly with pollen of other species, partly with that 

 of different mutants of 0. Lamarckiana, and the progeny counted 

 at the time of flowering. Six main groups were distinguished, cor- 

 responding to the old dimorphic races, and among these there were 

 about 8 per cent lata, 36 per cent scintillans, 1 1 per cent carta, 8 per 

 cent pallescens, 4 per cent liquida, and 4 per cent spathulata. Of 

 course, the figures are only approximate, since a large part of the 

 plants did not reach the period of flowering, but it should be noted 

 that the size of five of these groups lies between 4 and 1 1 per cent, 

 whereas only one, that of scintillans, is seen to differ widely. The 

 kind of pollen used for the crosses did not seem to have a notable 

 influence on the numerical relations between these divisions. 



Combining this result with our hypothesis concerning the dis- 

 tribution of the mutant characters among the six lateral chromo- 

 somes of the haploid nucleus (8), we find an exact correspondence 

 in the main points, but it seemed desirable to give further proofs 

 and to study these relations in more detail. 



We shall first give a survey of the groups concerned, and discuss 

 the individual deviations afterward. We call prototypes the forms 

 of the old races, which served for comparison. It should be remarked 

 that the mutants of our experiment cannot be expected to be exactly 

 the same as these prototypes, since their male parent was always 

 the velutina. In the prototypes the paternal half of the nucleus 

 is velutina only in the sesquiplex mutants oblonga, auricula, and 

 candicans, but \aeta in the dimorphic races (lata, scintillans, cana, 

 pallescens, liquida, and spathulata). This causes some differences, 

 especially in the leaves, which are broader for laeta and narrower 

 for velutina. In some instances this difference is striking, but in others 

 it is very small, as in the case of 0. cana. 



The isogamic mutants nanella and rubrinervis have not been 

 observed in our culture; no dwarfish or brittle specimens were seen. 

 Some plants were like Lamarckiana, and others constituted a new 

 type, which was almost like the parent species, but reached only 

 half its height. We call it 0. Lamarckiana mut. pulla (fig. 2). Two 

 plants repeated the features of 0. Lamarckiana mut. hamata, a 

 race cultivated since 1918, but as yet not published. The vast ma- 

 jority of the mutants showed more or less clearly the types of our 

 old dimorphic races. Table III gives the prototypes represented 

 and the numbers of individuals. 



From table III we see that the dimorphic races, with the excep- 



36 



