102 
SUMMARY OF CURRENT RESEARCHES RELATING TO 
reducing the aperture of your objective until all spectra are cut out, 
except those of the first order. The reasoning is as follows : With a 
small cone and an aperture sufficient to take in many orders of spectra 
on focal alteration, you obtain a series of changing images similar to 
those seen in a kaleidoscope. Without a priori conclusions you do not 
know your focus, consequently you cannot select the true diffraction 
ghosts from among the false diffraction ghosts. 
But the moment the aperture of the lens is contracted so that only 
the spectra of the first order are admitted, one image and one image only 
is possible. This image is certainly not a very good image of the 
structure, nevertheless it cannot be very dissimilar. 
In the case before us, instead of getting well-defined hexagons like 
those of a bee’s honeycomb, we have in place of them circular bright 
spots, spaced correctly and in arrangement precisely similar to the 
original. 
But it may be urged that all this only applies to diatom work, and 
has nothing whatever to do with ordinary microscopical objects. If you 
will pardon me for a moment I will endeavour to prove to you that it is 
of the highest importance with regard to almost every microscopical 
object, But first let me draw your attention, before leaving the Tri- 
ceratium, to a false diffraction ghost of the second degree (fig. 11). 
This picture is only possible when four orders of spectra are admitted. 
Here you will notice that each bar of the hexagon is broken up into 
three dots, and six spots with a central one are imaged in each areola- 
tion. This is a difficult one to photograph on account of the great 
brightness of the areolations, which accounts for the images in those 
parts being weak. To show that this is a subject not at all confined to 
diatomic structures, the next experiments will be performed on the eye 
of a fly. 
The spectra arising from this structure are identical with those 
from similar diatomic structures, only they are not so widely spread 
out, the intervals being 1/800 in. This proves that diffraction does not 
begin at 1/2500 in. I will first project the critical image (fig. 12) 
taken with a 3 4 cone X 165. The illuminating cone was now reduced, 
and the spectra, as in the next picture, allowed to pass into the objective 
(fig. 13). We now get the Eichorn intercostals. This shows that the 
diffraction theory has just as important a bearing in connection with a 
common entomological object as with a diatom. The next picture (fig. 
14) was taken with a large cone, but the aperture of the lens was 
reduced so that it should bear the same proportion to the eye of a fly as 
an oil-imm. of 1 *4 N.A. does to the P. angulatum. Here you will notice 
that the hexagon runs into a kind of square shape. A similar appearance 
can be obtained with a P. angulatum. 
The structure of the eye of the fly being very coarse it is possible 
to pick up the whole of the diffracted fan ; this, as seen at the back of 
the objective, is in itself such a beautiful object that I have endeavoured 
to produce it, but as yet without success. It is a beautiful star with 
hyperbolic edges, and is, as far as I am aware, unknown, nor figured 
anywhere. If this whole diffraction fan be admitted to the objective, 
then we get a false diffraction ghost of the third degree (fig. 15), and 
this is the last and most complicated ghost you can have. The founda- 
