1889.] MICROSCOPICAL JOURNAL. 225 



column during one of these pulsatory movements may be represented 

 by a curve, whose amplitude is constant for all temperatures with which 

 we have ordinarily to deal ; that the form and amplitude of this curve 

 are the same, whether the cycle of changes is completed within a few 

 minutes or in several houi's ; and, finally, that in accordance with the 

 facts of observation here pointed out, the uncertainty in the indications 

 of a mercurial thermometer is much greater in the case of slow changes 

 of temperature than in moderately rapid changes. 



" Referring only to the particular thermometers under investigation, 

 it is safe to say that random readings of a thermometer are in no case 

 reliable indications of the real temperature ; but if continuous readings 

 are taken at short and equal intervals of time until a cycle of changes 

 has been completed, the mean of the indications will indicate a close 

 approximation to the real temperature. This constitutes what I have 

 called a new method of obtaining the temperatme from the readings of 

 mercurial thermometers. Briefly, it may be called the method of read- 

 ing by cycles. 



" The systematic character of the deviations of one thermometer from 

 another is best shown by the comparison of the difference between the 

 readings at equal and regular intervals of time. 



" The advantages of the new method are obvious. If we wish to 

 determine the relation between a given thermometer and a standard 

 whose errors are known, we have only to make continuous compari- 

 sons under a moderately rapid change of temperature at various inter- 

 vals of time, and then take the mean of the differences for each com- 

 pleted cycle. In the case of my own standard, in which the period of 

 the cycle is y 2 ^ of a degree, there would be in a rise or fall of 10 de- 

 grees, occupying perhaps three hours, over 40 well-determined points 

 at which the relation between the two thermometers could be estab- 

 lished ; a result which would require several weeks of random observa- 

 tions made in the usual manner." 



In the discussion which followed Dr. Taylor said that while in Ger- 

 many he had noted cases where the glass of thermometers had been 

 left seven years to allow shrinking before the instrument was completed. 

 Other remarks were made suggesting that the variation of the glass had 

 much to do with the difference of degrees. 



Professor Rogers, in answering the discussion, pointed out the liability 

 to err in comparing thermometer readings. Professor Kellicott, Doc- 

 tors Lewis and Fell, and Mr. J. A. Miller, also discussed the theme. 



The next paper was by Professor S. A. and Mrs. Susannah Gage, on 

 " Staining and Permanent Preservation of Histological Elements, iso- 

 lated by means of Nitric Acid or Caustic Potash." 



The third paper of the afternoon was on " Microscopic Growth of 

 the Normal and Diseased Eye," by Dr. Lucien Howe, of Buffalo. The 

 doctor, who is one of the best authorities on all questions relating to 

 the eye, said that the outer coating of the eye was a common resting 

 place for bacteria, many of which are injurious and others that are harm- 

 less. He exhibited cultures, and also a patient suffering from eye 

 trouble. 



In the discussion which followed Dr. Taylor said it would be well to 

 appoint a committee to make investigations to ascertain whether or not 

 the bacillus in the human eye is similar to the bacillus in cases of pink- 



