On the other hand, we regard the development of the rudiments as depending upon motive forces or causes that are external to the germplasm of the ovum, but that none the less arise in orderly sequence throughout the course of the development. The causes we recognise are first, the continual changes in mutual relations that the cells undergo as they increase in number by division, and second, the influence of surrounding things upon the organism. One may group together as _centrifugal causes_ of the process of development the characters of the fertilised cells and the interrelations between the products of their divisions, and distinguish them from the _centripetal causes_, or motive forces that are provided by the action of surrounding things. None the less, it must be borne in mind that there is no sharp distinction between centrifugal and centripetal forces. On page 86 I showed how what is external in one stage of the process becomes internal in the succeeding stage. The external constantly is becoming internal, and the sum of the internal factors increases only at the expense of external factors. From the physiological point of view I regard the divergent differentiation of cells as a reaction of the organic material to unlike impelling forces--that is, to factors shown by experimental physiology to be actually present and to rule the building up of the organism. 'It were superfluous to detail,' as Naegeli says, 'how continually other forces external to the idioplasm, but belonging to the individual, influence the idioplasm; every cell, indeed, as it grows and divides, takes up a definite place in the growing whole, and finds itself in a peculiar combination of conditions of organisation.' 'Not only influences within the individual affect the idioplasm, as that may be altered by external influences, and so may be forced to grow in a new direction.' 'The influence of surroundings in determining which of the rudiments contained in the idioplasm shall achieve development is shown in the following example: it depends on their nutrition whether certain trees shall bear foliage or flowers; while in an unpropitious climate many plants refuse to bear flowers at all, but content themselves with vegetative reproduction.' This principle indicates the path along which explanation of the differentiation of cells is to be sought. Although in no single case is it yet possible to refer a known action to its appropriate cause--in other words, to show a definite stimulus producing a definite reaction upon the rudiment--this failure is not to be attributed to error in the principle. It is the natural result of the enormous difficulties besetting an attempt to understand the highly involved events of development. We can only ask whether or no our general principle is harmonious with the facts displayed in nature.