ZOOLOGY AND BOTANY. MICROSCOPY, ETC. 1021 



cells. In the finest ramifications the latter are at least as wide as, 

 and in most cases wider than, the sieve-tubes. In many dicotyledons, 

 especially those with bicollateral vascular bundles, the accompanying 

 cells of the finest ramifications of the veins have even a larger diameter 

 than in the leaf- stalk and stem. Koch's peripheral cells — the tran- 

 sitional cells of Fischer — are these broad accompanying cells. In all 

 collateral and bicollateral dicotyledons, with the sole exception of the 

 Cucurbitace^e, the finest ramifications of the bundles have a collateral 

 structure. In the sieve-portion which lies beneath the vascular 

 portion, imperfect sieve-tubes always occur along with the broad 

 accompanying cells, and the cambiform which is sometimes scarcely 

 distinguishable from them. These imperfect sieve-tubes are very 

 narrow, and have no distinct sieve-discs ; they contain little or 

 no protoplasm, and no nucleus, but sometimes mucilage ; frequently 

 they are filled only with a watery fluid. 



The blind ends of the veins in the lamina of the leaf are of two 

 kinds, principal and secondary. The secondary ends never contain 

 sieve-tubes, and in all collateral dicotyledons consist only of tracheids. 

 Among bicollateral dicotyledons, only the Cucurbitaceas have bicol- 

 lateral secondary ends to the veins ; the lower sieve-portion is in these 

 represented by a row of broad accompanying cells, the upper sieve- 

 portion by elongated cells containing but little protoplasm and no 

 nucleus. All other bicollateral dicotyledons agree with the collateral 

 in the structure of the secondary ends of the veins. The principal 

 ends of all collateral and bicollateral dictyledons have always a lower 

 complete sieve-portion, of the same composition as in the finest rami- 

 fications; and this always ceases before or along with the tracheids, 

 never after them. In all the princij)al ends are also the blind ends 

 of imperfect sieve-tubes. 



The author disputes the statement of Areschoug that in Ilex, 

 Tilia, and Buxus, the blind ends of sieve-tubes penetrate between the 

 cells of the loose parenchyma, independently of the veins ; the ob- 

 servation arises from a confusion with sclerenchymatous fibres. In 

 Buxus the blind ends consist only of sclerenchymatous fibres, which 

 have here assumed the function of tracheids, and must apparently be 

 regarded as elements for the conduction of water. In Buxus, Quercus, 

 Juglans, and Aristolochia, there are no blind ends of sieve-tubes in the 

 lamina of the leaf, and no principal ends. 



The sieve-portions of all dicotyledons examined always contain 

 sieve-tubes, accompanying cells, and cambiform. The cambiform, in 

 which the accompanying cells were formerly included, takes no part, 

 or only a subsidiary one, in the conduction of albuminoids ; its chief 

 function is probably to furnish the materials for the production of 

 albuminoids. The accompanying cells are the special scat of tho 

 formation of albuminoids, as is shown by their increase in size in tho 

 ultimate ramifications of the veins in the leaf. Fully formed sieve- 

 tubes take no part in tho production of albuminoids, but are tho 

 special organs for their conduction. 



A list of sixty-two species is appended, to all of which those remarks 

 apply, with tho limitations named. 



