104 FRESH-WATER ALG^ OF THE UNITED STATES. 



What part the nucleus has in the process just described I have never actually 

 demonstrated, but have little doubt but that it undergoes a division in the very 

 commencement, so that the nevi^ nucleus of each secondary frond is formed out of 

 one-half of the old one. 



In proportion as the form of the desmid becomes simpler, so do the peculiarities 

 of its cell multiplication become less. In those species which are simple cylindri- 

 cal cells, there appears to be nothing peculiar in the method of dividing, which, 

 however, always takes place through the centre of the cell, and subsequent growth 

 occurs, generally, only in the newly formed part. 



True sexual reproduction apparently does not take place as freely in this family 

 as the former process, for whilst I have seen hundreds of cells undergoing the 

 latter, it has not been my good fortune to meet with conjugating specimens on 

 more than two or three occasions. 



The process has, however, been studied very closely by De Bary, Braun, Hof- 

 meister, and others, and appears to consist generally in a rupture of the outer wall 

 of two cells and the protrusion of delicate processes from an inner, often newly 

 formed coat, with subsequent union of these, and consequently of the two cells, 

 and afterwards a condensation of the contents in the enlarged connecting passage. 

 The connecting passage between the fronds is really a sporangium in which the 

 spore is perfected, the contents of the cells finally condensing it into a firm globe 

 and secreting around themselves a thick coat. 



The after-history of this spore has been very successfully studied by M. Hof- 

 meister, whose observations were made upon Cosmarium tetraophthalmum, which 

 he watched conjugating and forming a sort of resting spore which was perfected 

 early in the month of July. This was composed of a thick outer coat and green 

 endochrome lying within as a distinct ball, nowhere in contact with the invest- 

 ing membranes. In three weeks' time this chlorophyllous protoplasm had divided 

 into ellipsoidal masses, or primordial cells, which soon surrounded themselves with 

 cellulose walls and became distinct free cells in the granular fluid which filled the 

 cavity of the original spore. In August, each of these masses was divided into two 

 and in the month of September the process was repeated, so that out of the original 

 endochrome eight strongly flattened primordial cells were produced. Division in 

 some specimens ceased here, and in others took place once more, so that by the fol- 

 lowing spring aU of the living Sporangia contained eight or sixteen green daughter- 

 cells, each of them discoid in outline with a strongly marked central notch. 

 These daughter-cells were finally set free by the solution of the spore wall, as Gos^ 

 maria of minute size, but agreeing in all other characters with the specific form to 

 which they belonged. 



According to Braun, in the larger, more or less lunate Closteria, conjugation 

 occurs in the following method : Two fronds approach one another in such a way 

 that they lie back to back. In the middle of each of them, there then appears an 

 annular line or trench reaching through the cell wall, and accompanied by a dis- 

 tinct separation of the endochrome into two halves. Whilst these changes have 

 been progressing there has also formed a new double wall at the position of the 

 trench, so that out of the two Closteria two pairs of separate equal cells have been 



