Their origin and growth has been most fully studied by Kleinenberg (No. 341) in Lum. trapezoides. They commence in this species shortly before the gastrula stage as two large cells on the surface of the blastoderm, which may be called mesoblasts. These cells lie one on each side of the median line at the hind end of the embryo. They soon travel inwards and become covered by the epiblast (fig. 141 A, _m´_), while on their inner and anterior side a row of small cells appears (_ms_). These rows of cells form the commencement of the mesoblastic bands, and in the succeeding stages they extend one on each side of the body (fig. 141 B, _ms_) till they reach the sides of the mouth. Their forward growth takes place mainly at the expense of the superjacent epiblast cells, but the two mesoblasts at their hinder extremities probably assist in their growth. Each mesoblastic band is at first composed of only a single row of cells, but soon becomes thicker, first of all in front, and becomes composed of two, three or more rows of cells abreast. From the above it is clear that the mesoblastic bands have, in L. trapezoides at any rate, in a large measure an epiblastic origin. [FIG. 141. THREE SECTIONS ILLUSTRATING THE DEVELOPMENT OF LUMBRICUS TRAPEZOIDES. (After Kleinenberg.) _ms._ mesoblastic band; _m´._ mesoblast; _al._ archenteron; _pp._ body cavity. A. Horizontal and longitudinal section of an embryo which is dividing into two embryos at the gastrula stage. It shews the mesoblasts and the mesoblastic bands proceeding from them. B. Transverse section shewing the two widely separated mesoblastic bands. C. Transverse section at a later stage shewing the mesoblastic bands which have approached the ventral line and developed a body cavity _pp._] At first the two bands end _in front_ at the sides of the mouth, but subsequently their front ends grow dorsalwards at the expense of the adjoining epiblast cells, and meet above the mouth, forming in this way a mesoblastic dorsal commissure. The mesoblastic bands soon travel from the lateral position, which they at first occupy, towards the ventral surface. They do not however meet ventrally for some time, but form two bands, one on each side of the median ventral line (fig. 141 C). The usual accounts of the origin and growth of the bands differ somewhat from the above. By Kowalevsky (No. 342) and Hatschek (No. 339) they are believed to increase in Lumbricus rubellus and Criodrilus entirely at the expense of the mesoblasts. Kowalevsky moreover holds that in L. rubellus the original mesoblasts spring from the hypoblast. In some forms, _e.g._ Lumbricus agricola, the mesoblasts are not present. In Euaxes the origin of the mesoblast bands is somewhat interesting as illustrating the relation of the Chætopod mesoblastic bands to the mesoblast of other forms. To render intelligible the origin of the mesoblast in this form, it is necessary to say a few words about the segmentation.