It has further been objected that, even if these various forms which Lamarck’s evening primrose throws off are true mutations, they ought not to be called new species, for they do not differ sufficiently from the parent species to deserve the name of new species. The reply to this criticism is that De Vries asserts that mutations produce new elementary species, which are not the same things as new species in the ordinary sense of the term. Most Linnæan species differ from one another to a far greater extent than do elementary species. It seems to us quite plain that new species arise, not by a single mutation, but by two or three successive mutations which occur in various parts of an organism. First arises a well-marked variety, by a single mutation. Subsequent mutations follow, so that a distinct race is produced. And, finally, fresh mutations occur, so that a new species is eventually produced. What De Vries calls an elementary species the majority of systematists would call a well-marked variety. We may take this opportunity of remarking that the definition of a species is one on which naturalists seem unable to agree. So vast is the field of biology, that now-a-days biologists are compelled to specialise to some extent. Thus we have botanists, ornithologists, those who devote themselves to the study of mammals, those who confine themselves to reptiles, or insects, or fishes, or crustaceans, or bacteria, etc. Now each class of systematists has its own particular criterion of what constitutes a species. Ornithologists do not seem very exacting. Most of them appear to consider a constant difference of colour sufficient for the formation into a species of the birds that display such a variation. Those who study reptiles, on the other hand, do not allow that a mere difference in colour is sufficient to promote its possessor to specific rank. Into these nice questions we cannot enter. For our purpose a species is a group of individuals that differ from all other individuals in displaying certain well-marked and tolerably constant characters, which they transmit to their offspring. Our contention, then, is that new species, in the ordinarily-accepted use of the term, do not arise as a rule by one sudden bound (although they may sometimes do so), but are the result of the accumulation of several mutations or discontinuous variations. Some of these mutations are exceedingly well marked, while others are so small as to be indistinguishable from the more extreme fluctuating variations. Before passing on to consider some cases of well-marked mutations which have occurred among animals and plants, we should like to take this opportunity of pointing out that as regards experiments in evolution the botanist is far more favourably situated than the zoologist.