MONOGRAPH OF THE LABOULBENI ACEvE. 373 



are remarkably variable. Tbe born-like appendage, wbich cbaracterizes tbc perithccia of the 

 species first described, does not prove of generic value, since it is absent in four forms. The only 

 genus which might possibly be mistaken for it, however, is Sphalcromyces ; but the presence in this 

 case of flask-shaped antheridia on the branches of the appendage, not to mention its well-developed 

 perithecial stalk-cell, distinguish it without difficulty. Its greatest peculiarity lies in the character 

 and development of the perithecium, which differs from the more typical forms in important 

 points ; and although I have been unable to obtain young material which illustrates all the early 

 stages, it appears in Cmirabilis to be as follows : The young individual consists of a simple scries 

 of superposed cells, the distal of which begin to branch at an early stage ; and the first indica- 

 tion of the formation of the perithecium consists in the division by a transverse septum of one 

 of the cells of this series. Of the two cells resulting from this division, the upper grows out- 

 ward and upward, its outgrowth becoming a finger-like projection (Plate XXIV, fig. 4, at the 

 left) that later elongates to form the filamentous trichogyne. This outgrowth, in the stage rep- 

 resented by the figure, has become divided by two transverse partitions into three cells ; the upper 

 being the young trichogyne, the cell below it the trichophoric cell ; while below this, not projecting 

 from the axis of the plant, is the smaller carpogenic cell. At the right of the carpogenic cell a 

 vertical septum has divided the remainder of the original upper cell into two parts, an outer and 

 an axile cell ; but in other species, like C. contortus (Plate XXV, fig. 9), this division seems to 

 be absent. The lower of the two original cells has, in the mean time, been divided by a vertical 

 septum into two cells, one of which (Plate XXIV, fig. 4, x) has also begun to grow outward and 

 upward. As far as can be determined, this cell (x) by further growth and division produces 

 only one of the series of wall- and canal-cells ; while the cell previously referred to as the axile 

 cell, lying behind the carpogenic cell, gives rise to the three other series. Having been unable 

 to separate the cells in this region and ascertain their protoplasmic connections with the cells of 

 the perithecium, it is uncertain whether the above statement is entirely correct in so far as con- 

 cerns the number of wall-cells originating from each of the cells mentioned ; but that they give 

 rise to the outer and inner series of perithecial cells cannot be doubted. The latter grow up 

 around and beyond the carpogenic and trichophoric cells, and by the successive division of their 

 terminal members form the main bulk of the perithecium. The base of the trichogyne is thus 

 left behind, as it were, in the angle between the perithecium and the appendage (Plate XXIV, 

 fig. 5, and Plate XXV, fig. 17), where it often persists for some time. While in other genera, 

 as a rule, only one or two divisions of the perithecial cells takes place above the insertion of the 

 trichogyne, in the present instance the growth beyond it may be almost indefinite ; as in C. ros- 

 tratus, for example, or C. filiformis, in each of which (Plate XXIV, fig. 15 and 18) the perithe- 

 cium is extraordinarily developed. Although the general process by which the perithecium 

 originates is thus very similar to that which has been described as characteristic of Stigma- 

 tomyces or Laboulbenia, it will be observed that its basal cells are differently arranged, and 

 originate in a different manner. Whether it is possible to homologize the typical stalk-cell and 

 secondary stalk-cell with any of those present in this case I do not feel able to determine. The 

 perithecial cells of the inner series differ from those of other genera from the fact that they seem 

 to divide quite independently of, and less frequently than, the wall-cells ; as may be seen in 

 fig. 8, Plate XXIV, the septa of the wall-cells (w c) and of the canal-cell (c c) in this instance 

 being by no means coincident. As in other cases, the canal-cells arc entirely destroyed by the 



