244 ADJOUENED DISCUSSION IN LONDON 



are almost perfectly transparent, and yet metals are the most opaque 

 of substances. Latterly I have been endeavouring to photograph by 

 means of ultra-violet light some exceedingly small organisms, some 

 of which are beyond the limits of resolution, and the difficulty has 

 been that with any wave-length I have at present available, the 

 organism is transparent. The radiations pass completely through, 

 and I am unable to get an image of any description whatever. So 

 that to say that the limitations of the work are largely governed by 

 the opacity of small bodies is not in accordance with practical 

 •experience or theoretical expectations. It may possibly arise if we 

 use radiations of much shorter wave-length than those at present 

 available, but in that case we shall be working with a microscope in 

 vacuo, and I do not think it is a point which is likely to arise in 

 practice for some time to come, although it may, and probably will, 

 arise at a later stage. 



Mr. L. C. Martin: I was interested in the description of the 

 Hartridge test for the microscope objective, but I should like to say 

 that it is not often, I believe, that a man testing a microscope objec- 

 tive wishes to know the aberration to any great accuracy, but rather 

 whether the microscope objective is sufficiently good for the purpose. 

 Therefore a somewhat easier quantitative test is to be desired. At 

 the present time I have been doing a certain amount of work as a 

 sort of preliminary study of the star test, and I thiiik that possibly 

 the so-called Rayleigh condition of less than one quarter wave-length 

 a speedy test of the aberrations of a microscope objective. 



Professor Conrady remarks in his paper that the fulfilment of 

 the so-called Rayleigh condition of less than one quarter wave-length 

 difference of optical paths between paraxial and marginal rays iu 

 good telescope and microscope objective, has been demonstrated by 

 the Hilger interferometer. It is easy to understand that, imagining 

 a perfectly spherical mirror in the interferometer and a means of 

 controlling the position of such a surface to correspond with any 

 particular focus of the test lens, such a perfect demonstration could 

 be given. It is not easy to understand, however, when we consider 

 that the errors of the surface of a mirror, which may be of the order 



<5i — or even more, are doubly important in such a case, and that 



the position of the test focus has to be obtained by trial. It is only 

 when we consider a fact which was hinted at by Lord Rayleigh in 

 1879, and worked out by Professor Conrady in his paper on Star 

 Discs, viz., that the effects of spherical aberration can often be 

 countered very completely by changes of focus (or in mathematical 

 language that we can partly balance the terms of the fourth and 

 higher orders in the aberration expression by a change of the co- 

 efficient of the second order), that we can realise that the indications 

 of the interferometer are trustworthy even to the extent previously 

 indicated. It is necessary to bear in mind, however, that there is 

 nothing magically sensitive in the interferometer tests as compared 

 with star tests, for example, if these are performed with the maxi- 

 mum of care. Those who expect them to give tremendously sensi- 

 tive results far excelling all other tests are doomed to disappointment. 



