110 M. Miiller on the Development of the Lycopodiacese. 



of the cells of the terminal bud remains yet to be traced. I 

 therefore next consider — 



2. The roots of the branch. These are solid round bodies, al- 

 most perfectly cylindrical, which always split, like the axis, dicho- 

 tomously when they come in contact with the ground, penetrate 

 it, and attach the plant more firmly to it. When growing in pots 

 however, as in our green-houses, and not long enough to reach 

 the earth, they do not divide. 



Interiorly they are made up of the same kind of cellular tissue 

 as the axis of the branch ; but the cells are firmer. They never 

 contain more than one vascular bundle, and also differ from the 

 axis in the fact that this bundle is not attached to the cortical 

 layer by those tubular cells, whereby a large empty space is formed 

 in the centre of the axis of the branch. 



The most remarkable point about them is their extremely re- 

 gular occurrence upon the axis of the branch. They always make 

 their appearance in the situation where the axis splits into two, 

 and in fact immediately in the axil of the last or penultimate leaf 

 of the branch (PL II. fig. 17), which leaf is always a folium inter- 

 medium ; the root consequently is always developed upon the upper 

 side of the axis. Schleiden (in his Grundziige, ed. 1. part ii. 80) 

 expressly says of the roots, that they proceed from the under side. 

 He has not mentioned the species in which he observed this, and 

 therefore I am still in doubt as to the truth of the statement. 



The rootlet appears at first as a little papilla upon the axis of 

 the branch. This soon developes into a conical projection, and 

 as soon as it has attained a certain length, which is not very con- 

 siderable, the little cone curves downwards. This structure begins 

 to be developed almost in the earliest stage of the growth of the 

 axis of the branch, and the rootlet itself is found in that situation 

 where the organs of reproduction are produced ; a condition to 

 which I shall have again to refer hereafter. Every joint of the 

 branch consequently possesses a root at its base. A successful 

 transverse section exhibits the history of development in the in- 

 terior ; at the particular point of the axis, where the root is to be 

 formed, the tubular cells round the vascular bundle partly dis- 

 appear ; with the disappearance of these the vacant spaces also 

 vanish, and the cortical parenchyma now immediately invests 

 the central vascular bundle. On the upper side of the axis 

 alone, consequently directly at the spot where the rootlet subse- 

 quently becomes visible, some tubular cells still occur, which 

 however are of small importance compared with the former of the 

 axis of the branch (PL IV. fig. 1). Indeed they are really only 

 parenchymatous cells elongated upwards. Since these cells are 

 wanting and the parenchyma is thus brought close to the vas- 

 cular bundle, the nutrient fluids must naturally become con- 



