ENGELMANN — NORTH AM. SPECIES OF JUNCUS. 



431 



with such different length of tail that only the absence of 

 any other diagnostic characters can induce us to consider 

 them as belonging to one and the same species. Much less 

 can generic distinction be based upon this character, as was 

 done by Desvaux, who comprised in his genus Marsipposp&r- 

 mum all Junci with tailed seeds. Even E. Meyer's (in Synop- 

 sis Juncorum, 1822, and in Ledebour's Flora Rossica, 1853) 

 separation of the species with tailless seeds as a second section 

 is unnatural, as not only tail-seeded kinds are found in all the 

 great groups, but also species with intermediate seeds exist, 

 which it would be difficult enough to place properly. R. Brown 

 (Prod. Nov. Holl., p. 258) settles the whole question in the 

 following pithy sentence: Nee secernendm sunt ew quce 

 seminibus gaudent scobiformibus, testa silicit, quaz in pluri- 

 bus utrinque laxa, in hisce valde elongata. 



The size of the seed varies from 0.1 to 2.0 lines in length, 

 it mostly ranges between 0.2 and 0.3 lines, and rarely reaches 

 0.4 lines; the tailed seeds are usually larger than the others, 

 averaging from 0.5 to 2.0 lines in length ; even without the 

 appendage, J. triftdus has the bodies of the seeds of 0.5, /. 

 castaneus of 0.5-0.6, and J. sty gins of 0.7-0.8 line in length. 

 The delicate markings of the seeds are so various, and in 

 the same species so constant, that it will be useful to dwell a 

 little longer on them. Their surface appears never quite, 

 and rarely nearly, smooth, when magnified fifty or sixty times. 

 We can almost always discover longitudinal ribs, more or less 

 close together, from 8 or 12 to 30 or 40 or more around the 

 seed; as it is difficult to count the ribs all around these^ small 

 bodies, and as an approximate designation is quite sufficient, 

 only the number visible on one side may be counted. These 

 ribs are very marked, sharply elevated, in /. marginatus {sem- 

 ina costata), or they are reduced to more delicate lines in J. 

 Canadensis and most tail-seeded species {semina multico- 

 stata and striato -costata). These ribs or lines are usually 

 united by very delicate transverse lines (lineolce), when such 

 seeds may be termed costato-lineolata, or by fewer, more 

 prominent cross-bars, semina costato-reticulata. 



When the ribs are fewer and wider apart, and united by 

 transverse ridges so as to form somewhat rectangular meshes, 

 I call the seeds semina reticulata. The area of these meshes 

 is sometimes quite smooth (J. militaris), or crossed with very 

 tew transverse or longitudinal lines {J. scirpoides) — semina 

 areis Icevibus reticulata — or it is distinctly marked by numer- 

 ous delicate transverse lines, sometimes, also, with one or 

 two perpendicular lines: semina areis lineolatis reticulata. 

 In very few instances I find an irregular and indistinct retic- 

 ulation : semina irregulariter sub-reticulata. 



A large number of Junci exhibit a very delicate but regular 

 transverse reticulation without (in fully ripe seeds) very 



