370 THE BIOLOGY OF BIRDS 



are often very important for the individual, but it is uncertain 

 whether they can be entailed as such on the race. 



Now when we subtract from the total of observed 

 differences all that we can interpret as modifications, and 

 when we also put aside differences due to age and sex, the 

 remainder consists of inborn variations. These are due to 

 novel permutations and combinations in the germinal 

 material that developed into the organisms we are studying. 

 Some novel arrangement or disturbance in the intimate 

 architecture of the egg-cell and the sperm-cell, expresses 

 itself in development as a variation or new departure — 

 endogenous not exogenous, an outcome not an imprint. 

 There can be little doubt that these germinal variations 

 furnish by far the greater part of the raw material of evolution. 



New departures or variations in the wide sense may be 

 conveniently ranked in two groups : (a) continuous, quantita- 

 tive variations, such as a slightly longer pinion or a slightly 

 stronger bill, and (b) discontinuous, more qualitative varia- 

 tions, such as the loss or addition of a wing-feather, the gain 

 of a crest or a spur. The former may be spoken of as inbor?i 

 fluctuations, a little more of this and a Uttle less of that, but 

 they must be kept apart from similar changes {acquired 

 modifications) which are due to peculiarities of individual 

 nurture. The discontinuous variations, which occur 

 brusquely, without grading into one another or into the 

 characters of the parents, may be called mutatiofis. What is 

 inborn or germinal in origin is always heritable, unless it be 

 an " impossible " variation like sterility ; but it does not 

 follow that it is always " transmitted " or entailed. As a 

 matter of fact the forthcoming evidence of entailment is 

 stronger for mutations than for fluctuations, but this inquiry 

 is still very young. On general grounds — the continuity 

 of the germ-plasm — a variation in the germ-cells is likely to 

 be continued in the inheritance unless it be of a disintegrative 

 character, such as some pathological disturbance. 



So far then, the terminology runs : — 



IP ■ 1 V9 ■ t' q / Innate mutations. 

 \ Innate fluctuations. 

 Somatic Modifications 



