232 MendeVs Laws of Alternative Inheritance in Peas 
3. The colour of the seed-coat (grey, grey-brown, or leather-coloured, with or 
without violet dots, associated with purple-violet flowers and red colouring round 
leaf-axils, dominant ; white, associated with white flowers, recessive). 
4. The shape of the i'ipe pod (not constricted, smooth, dominant ; constricted 
between the seeds, recessive). 
5. The colour of the unripe pod (green, associated with green stem, mid-ribs of 
leaves, and calyx, dominant ; yellow, associated with similar colouring of stem, 
mid-ribs and calyx, recessive). 
6. Tlie distribution of the flowers (scattered along the axis, dominant; gathered 
into a short umbel-like cluster, at the extremity of the axis, recessive). 
7. The lengtJi of the main stem (tall, dominant; short, recessive). 
The only qualifications Mendel offers, in applying his general statements to 
these very varied characters, are (1) that the violet dots on the seed-coat are often 
more numerous and larger in hybrids than in pure-bred forms, and (2) the 
observation that the mere fact of hybridisation produces an increase in the size of 
the vegetative organs, so that hybrid plants are often taller than either of their 
parents, an observation made previously by Knight (JSIo. 18) for peas, and by many 
later naturalists for peas and other plants (see the summary of the evidence by 
Darwin, No. 9, Vol. u. Chap. 17, and in addition to the authorities there cited, 
Naudin, No. 22. For extensive observations on Peas, see Tschermak, Nos. 27 
and 28). 
It is clearly important to test these remarkable statements by a careful study 
of the numerical results, and by the application of such tests as may be possible. 
It seems to me that by neglecting these precautions some writers have been led to 
overlook the wonderfully consistent way in which Mendel's results agree with his 
theory, saying that his numbers " are not large enough to give really smooth 
results," and while they thus unwillingly do rather less than justice to Mendel's 
own work, at the same time they accept results which seem really inconsistent 
with those obtained by Mendel as proof that his statements are applicable to a 
wider range of cases than those he actually observed. 
Mendel's observations fall into two groups ; the first group relates to a series of 
cross-fertilisations, in which each pair of races crossed differed in only one of the 
seven sets of characters dealt with ; the second group contains observations on 
races which differed in two or more sets of characters. 
The observations of the first group are more numerous than those of the 
second, and may be considered first. 
The seven sets of observations, showing the dominance of one character in the 
first hybrid generation, must of course rest upon Mendel's statement, which I think 
no one who reads his paper will find the slightest difficulty in accepting. The 
behaviour of hybrids of the second generation can be tested from the numbers 
