In fig. 4 _a_, I have given a drawing of the two chains of glands, but with only those ducts figured which proceeded from the last-formed pair of glands. The specimen here drawn was old; and it is rare to find the structure of the ducts so simple. From both glands[72] a neck or thick duct arises, which soon bifurcates; one branch runs direct into the circumferential duct, and the other (_t_) bifurcates again; of the latter, one branch unites with its fellow from the opposite gland, and then forming a single duct (_t′_) enters, as do the two other branches, the circumferential duct. Thus, into the latter, five main ducts enter: the position of their points of entrance, with respect to the shell, varies considerably; but I think the five points tend to face the middle of the rostrum, and middle of the two lateral compartments on each side. In some other specimens, in which the ducts were nearly as simple, I observed that the neck or main duct at once divided into three branches, instead of into two, with one soon bifurcating; and on one side a rudimentary branch or spur was given off (above _t_), indicating a tendency to an additional bifurcation. In the later-formed glands, the ducts proceed only from the outer sides and form the ends of the glands furthest from the centre; but in the earlier-formed and smaller glands of the same individual, other ducts proceed from the inner sides, where in the older glands the spurs (_m_) are situated: moreover, in the younger glands, all the ducts bifurcate much oftener (how often I was not able to ascertain), before entering the circumferential duct; many of the branches, however, terminating in spur-like points. Now if we imagine twenty or thirty repetitions of the ducts given in fig. 4 _a_, (independently of the greater complication of the ducts of the younger glands), each a very little smaller than the other, and placed, with the main branches parallel, one over and within the other, we shall gain some insight into the wonderfully complicated structure of the cementing apparatus in this and many other species of Balanus. [72] It should be observed that fig. 4 _b_ ought to have been drawn with its present upper end downwards, to make it correspond in position with fig. 4 _a_.