ON CROSSES AVITH DISSIMILAR HEREDITY. 
251 
occur, the cross type remaining always the same as it was immediately 
after the cross. 
Other crosses in the subgenus Onagra with the older species behaved 
precisely similarly. In 1894 I crossed CEnotliera Lamarckiana ? with 
(27. biennis S . The cross-type was uniform, and was maintained without 
separation in subsequent sowings. CE. Lamarckiana x CE. muricata gave 
same results, &c. 
Uniformity of the first generation is the rule for the Mendelian 
crosses, independently of secondary characters which arise through 
unusual nourishment of the seed, &c. For the false crosses, however, it 
is by no means to be expected, as we have already learned from Mendel's 
Hieracitim crosses cited above. 
The new species of CEnotliera, which originated in my trial garden, 
have afforded me splendid material for the study of this non- uniformity of 
the first generation. I observed this non- uniformity clearly for the first 
time in 1895 in a crossing made in 1894 with CE. lata, and have annually 
since then effected more and more of such unequal heritage crosses. The 
genus CEnotliera is particularly fitted for such purposes, since the crossing 
can be conveniently effected, each of them yields a sufficiency of seeds, 
and especially as the various types are often easily and distinctly to be 
distinguished in the quite young plants. 
For the first example I select a crossing between CEnotliera 
Lamarckiana •$ and CE. nanella ^ .* In the summer of 1898 I have 
effected crosses between these two plants, and sowed the resulting seed 
separately. The progeny, i.e. the first generation of crosses arising from 
these seeds, displayed always two types, that of the father and that of the 
mother, but in very varying ratios. I have counted the separate crops, 
and found the following percentages of CE. nanella :— 1, 2, 3, 3, 4, 5, 5, 
7, 7, 7, 9, 9, 11, 13, 13, 13, 15, 16, 17, 17, 17, 19, 19, 19, 20, 21, 21, 21, 
22, 23, 23, 24, 24, 26, 26, 28, 35, 39, 48. Or, grouping the separate 
crosses together — 
Percentage wa?ie^Za : 1-5, 5-15, 15-25, 25-35, 35-50 
Crossings : 7 10 16 4 2 
showing an average of about 17 per cent. 
In previous and subsequent years I have made about an equally 
large number of crosses between these two plants. They yielded very 
rarely 0 per cent., and so far never over 50 per cent, nanella. 
In the subsequent generations these crosses when self-fertilised were 
of both types and constant. The second generation consisted therefore of 
(E. nanella alone from the CE. nanella resulting from the first cross, and 
of CE. Lamarckiana alone from this type simultaneously raised. So far, 
however, as the grandmother used for the crossing was mutable, so were 
the offspring derived from the cross. 
So likewise was it with fifty crosses effected in 1898 between CEnotliera 
lata ? and CE. Lamarckiana ^ . The progeny displayed the two parental 
types, and CE. lata resulted thus : — 
Percentage lata: 1-5, 5-15, 15-25, 25-35, 35-45 
Crossings: 2 6 24 11 7 
showing an average of about 23 per cent. lata. 
* Compare Botanisch Jaarhoek DocloncBa, 1897, p. 73 ; and Coniptes Rendiis dc 
VAcad. des. Sc. Paris, Oct. 1, 1900. 
