EPITHELIUM.
their being flattened. The serous layer of the germinal membrane also cannot well be considered to be epithelium, although it has the same structure, and yet it 1s difficult to give a definition of it which shall not comprise these structures. We shall not, however, enter upon this contention about mere terms, but proceed to the consideration of the structure of the epithelium.
The simplest form of epithelium is that of the round cells furnished with a nucleus which lies upon the inner surface of their wall, and encloses one or two nucleoli. When in con- nexion they assume a polyhedral form, but their free surface usually projects in the form of a segment of a sphere. Such is the appearance presented by the epithelium in many situations ; I instance only that of the branchial rays of the fish by way of illustration. The cells are usually smaller and more granulous in mammal; but in the lower animals and in the foetal stage of mammalia they are, in general, larger, smoother, and sometimes so transparent as to be visible by a subdued light only. I once had an excellent opportunity of observing the epithelium upon the mucous membrane of the stomach of a foetal sheep, and its perfect resemblance to the parenchy- matous cellular tissue of plants. A minutely granulous deposit may often be observed in the interior of the transparent epithelial cells; in those of the branchial rays of the fish, for instance, it appears to be formed in the neighbourhood of the nucleus. According to Henle, two nuclei never occur in an epithelial cell in mammalia; but I have several times observed that number in the external covering of the tadpole, and on one occasion | remarked that a perfectly developed epithelial cell furnished with a nucleus was enclosed within a larger cell. Changes in form from this rudimentary globular shape occur in the epithelial cells in two different manners; they either become flattened into tables, or prolonged into cylinders. The flattening out into tables takes place in such a manner that the nucleus forms the centre of one surface, as in the blood-corpuscle. I have observed the stages of transition from the globular to the tabular form in the epithelium of the external covering of the tadpole, which occasionally presented hexagonal flat columns or tables, the thickness of which was about equal to one third of their breadth. The thickness is so very slight in proportion to the breadth in the completely flattened epithelial cells, that
