276 ANIMAL MORPHOLOGY 



ficance of chromatin reduction and of its unexpected connections 

 with the new phenomena of the first embryological phases (Borel, 

 Moore, and Farmer). 



The idea of phagocytosis, studied by Haeckel in the biology of 

 the protozoa, by Rouget in the examination of the leucocytes of the 

 blood, increased in value and importantly developed by Metchnikoff, 

 who made many applications of it in the domain of pathology, has 

 come by a most fortunate turn of events to elucidate certain of the 

 most obscure of the morphological phenomena of embryology, the 

 coenogenetic processes of ovogenesis and of metamorphosis. 



For a long time the introduction of the mathematical sciences in the 

 domain of morphology has been regarded with suspicion; it seemed 

 dangerous indeed to wish to bind by very simple formulae facts so 

 complex as those studied by zoologists and botanists. 



Little by little, however, the necessity of determining by precise 

 measures the extent of variations due to primary factors and of 

 seeking to find the laws of these variations has made itself felt. Among 

 the first Delboeuf attempted not without success the application of 

 algebra to the problem of the formation of races. But it is especially 

 to Galton and to his school that the most important of the works of 

 mathematical biology and biometry are due. 



Whatever be the character to which we give our attention, if we 

 consider a great number of specimens of a given species we recognize 

 that the individual variations (continuous or fluctuating variations) 

 of this character, numerically expressed, do not exceed two extreme 

 limits which are reached by a very small number of the individuals. 

 Between these two extremes there is a constant mean variation with 

 the greatest number of observed specimens. It results, that if we 

 take as abscissa the lines which represent the extent of the fluctua- 

 tions and as ordinates the distances corresponding to the number 

 of individuals which present a certain fluctuation, we obtain a curve 

 which Quetelet called a binomial and which is in reality only a curve 

 of probable error. We also often give to these curves the name 

 Galton's curves, because of the very extended use that this eminent 

 biologist made of them in the study of the question of heredity. 



By artificial selection breeders and horticulturists succeed in dis- 

 placing more or less rapidly the apices of the Galtonian curves and 

 in directing fluctuation in the way they desire. Natural selection 

 does not operate otherwise for the modification of the form of species 

 and it is to this action that Darwin attributed in great part the trans- 

 formation of species. 



Wallace, more especially, considered selection as the sole factor 

 determining the evolution of living beings. 



It was reserved for Hugo de Vries to show through long and delicate 

 cultural experiments, the exaggeration into which the immoderate 



