1898.] MICEOSCOPICAL JOUHl^AL. 115 



that can really be learnt from it, is tliat there is probably 

 (for this is subject to possible delusion from the false 

 intercostals above mentioned) some periodic difference of 

 structure in the object similar in dimentional intervals 

 to " lines " shown : in regard to " spots " this is more 

 uncertain, since these are often produced by false diffrac- 

 tion-fringes from any long line which may cross the true 

 ones. That the lines are lines, or that the " pattern " is 

 so geometrical as appears, is in the highest degree im- 

 probable. That the "spectrum" theory and method so 

 long retained exclusive predominance, is because atten- 

 tion has been so concentrated upon either gratings or 

 diatoms of known periodicity in structure, but which 

 only represent to a very small extent indeed any serious 

 kind of investigation. 



18. It appears that in microscopy we have to deal 

 with two characteristics of an image, which often are only 

 to a limited extent compatible ; that we have at command 

 two methods of illumination which respectively promote 

 more especially each of such characteristics ; and that in 

 most cases our problem is so to combine and balance 

 these two methods as to produce the best result. 

 Fidelity of contour will be secured in proportion as we are 

 able to obtain our image by heterogeneous illumination, 

 approximating the object to a self-luminous condition. 

 But this method may prove utterly unable to give us 

 contrast, which we may therefore be compelled to increase 

 by using to a greater or less (even to a very large) extent 

 plane-wave illumination, at the expense, however, of 

 some greater or less degree of infidelity in contour. 

 Thus an opaque subject, even of much minuteness, may 

 be best shown by ground-glass illumination, or a very 

 wide cone ; while a diatom, unless in a very dense me- 

 dium, or dry in air, may require narrow pencils of 

 approximately plane waves. It is interesting to observe 

 that there is thus a great degree of practical truth in 



