140 
Wallich, on the Diatom-valve. 
portion of the valve^ and fracture takes place through their 
centres whilst in Pleurosigma, the intra-linear spaces, which 
are held, but erroneously, to be equivalent to the hexagonal 
spots, constitute the strongest portion of the valve, and 
fracture never takes place through them, save when undue 
force happens to be employed. That the error in this case 
originates in a misconceived analogy, is evident from the 
subjoined quotation: — ^^The valve is thinner and weaker at 
the parts occupied by the dots ; so that the line of fracture 
corresponds to these parts/^ And again — "In those 
{Pleurosigmas) requiring the use of oblique light and stops, 
the line of fracture also corresponds to the rows of dots, 
provided the light is equally oblique on all sides; and the 
same appearances are presented by the dots in both cases, 
beginning with those in which they are large (as Isthmia), to 
those of moderate size (as in the species of Coscinodiscus) , 
down to those in which they are extremely minute (as in 
Gyrosigmaj he). Moreover, analogy affords us very strong 
confirmatory ground j for the Diatomacese form a very 
natural family, and if the dots are depressions in some 
genera, we might expect them to be so in others. Had 
these dots (in Gyrosigma) represented elevations, the valve 
would have been stronger at those points. (^ Micrographic 
Dict.,^ new ed., p. 34.) 
If we take into consideration the outline of the more 
marked discoidal forms, for instance, Triceratium and Cosci- 
nodiscus, and contrast it with that of the Naviculoid group, 
such as Navicula or Pleurosigma, it appears to me that we 
might naturally expect to find the markings in the two types 
exhibiting some distinct relation to the outline. Now, in the 
genera which exhibit the honeycomb structure — that is, where 
we find the appearance of a number of little hollow cylinders, 
of considerable relative depth, and open outwardly in so far 
as the siliceous wall is concerned — the conversion of elemen- 
tary circular cavities into hexagons is exactly what would 
result from pressure exerted equally in every direction. In a 
small but well-marked species of Coscinodiscus, C. nitidus, 
Greg., the markings are always circular, the distance between 
them being too great to admit of their shape being modified 
by the pressure of each cellule upon those adjoining it. The 
same holds good of a small Triceratium, closely allied to Mr. 
BrightwelFs T. punctatum. 
In Isthmia we meet with shallow cellules, or rather depres- 
sions, varying in different frustules, in different gatherings, 
and in different parts of the same valve, from irregular circles, 
to irregular hexagons, parallelograms, and pentagons, accord- 
