792 SUMMAKY OF OURKENT RESEARCHES RELATING TO 



species. In his ' Synopsis of tlie British Diatomacese,' vol. ii. p. xxiv. 

 of the introduction, Smith expresses himself thus : " Whether the 

 strife .... are mutually parallel or radiate in their arrangement ; 

 reach the median line or are absent from a greater or less portion 

 of the valvular surface ; — whether the cellules themselves are ar- 

 ranged in squares or disposed in quincunx ; and the striee in their 

 transverse direction are consequently parallel or oblique in refer- 

 ence to the margin or the median line — the relative distances of the 

 striaB and their greater or less distinctness — all these are features 

 which may safely be regarded as of specific importance . . . , 

 though subject to slight modifications arising from the accidents 

 of locality and age, &c." However great this authority may be, 

 I have preferi'ed to rely upon experience, which is in fact the 

 sole supreme judge and without appeal in such a matter. I think 

 that two forms found in different localities and at different times, 

 however great the affinity which they may present inter se in their 

 characters, may form two types closely bordering ui^on one another, 

 but nevertheless independent and distinct, and destined to produce 

 their own form through an indefinite number of generations. We 

 see this fact hajipen in many other orders of organisms. Whether 

 these are better called varieties than species, seems to me wholly 

 indifterent, the distinction between the two words depending on the 

 different value which is commonly given to them, a value and signifi- 

 cation considerably elastic — provided always that we recognize the 

 permanence of the form in the successive generations; and this exact 

 permanence of form in the smallest peculiarity of structure in the 

 Uiatomaceai is demonstrated as a constant law by my study of those 

 of the carboniferous e2)0ch. 



However, when I happened in a certain locality to find a collec- 

 tion, composed of myriads of frustules evidently belonging to a 

 single species, and forming a single family,* I detei'mined at once to 

 make a most attentive examination, feeling assured that in such 

 circumstances only could the most conclusive elements be obtained for 

 recognizing the extent of the possible modification of the typical form. 

 These circumstances are not at all frequent, although they are some- 

 times found ; I therefore recall with regret that in the early times of 

 my studies on the Diatomacefe, being desirous of obtaining new types, 

 I was vexed at finding gatherings composed of one or two forms only, 

 all which I rejected, not knowing then the profit which I might draw 

 from them. In gatherings of this kind formed of masses of frustules 

 belonging to a single sj)ecies, we find a series of forms varying in 

 size and presenting sometimes notable variations in outline. But in 

 reviewing the different dimensions and oxitlines of the frustules, we 

 recognize after attentive examination that amongst these frustules 

 there is a complete gradation, so that in passing from one to the 

 other by insensible degrees we prove the identity of the extreme 

 terms of the scries. Moreover, when we see at the same time that the 

 disposition of the striae on the valves (rows of granules or any pecu- 



• * Produced by a aiugle sporaugial frustule. — Ed. 



