ON TUB ZOOLOGICAL STATION AT NAPLES. 99 



per cent, alcohol. Between these two extremes lies the desirable mean. 

 Theoretical l}"- the change, in passing from the bottom of the reservoir to 

 (he top, should be perfectly gradual and uniform from per cent, to 

 :10 per cent. This would justify the name which I have given the 

 instrument, ' Tiie Differentiator,' and, in fact, the approximation to the 

 theoretical perfection may be so close a one that I retain the name first 

 suggested by the mathematical analogy. 



Even before the reservoir is full the pressure causes a slow flow 

 through the filter, and, drop by drop, the whole contents filter through, 

 and the object is thereby slowly transferred to 30 per cent, alcohol. 

 Hurtful diffusion currents are completely avoided. 



The rate of flow is governed by the size of the capillary part, the 

 tightness of the plugs, and the inclination of the instrument. The first 

 two of these factors are approximately controllable, the third exacfhj so. 

 The rate of flow, and therefore the period of time occupied by the change, 

 may consequently be accurately regulated. I find it best to prolong the 

 period to ten hours in the following cases : Sublimate to 30 per cent. ; 

 30 per cent, to borax carmine ; borax carmine to 50 per cent. ; 50 per 

 cent, to 70 per cent. ; 70 per cent, to 90 per cent. ; 90 per cent, to abso- 

 lute ; absolute to ^ turpentine ; ^ turpentine to turpentine ; turpentine 

 to dilute balsam. It is well to bear in mind that at length the flow is 

 slower than at first, owing to a decrease of pressure, though the diS"erence 

 is not so great as might be expected (on account of the influence of the 

 capillary tube). This forethought will lead to the insertion of rather 

 greater quantities of the heavier fluids than would otherwise be the case. 

 The next change, from 30 per cent, alcohol to borax carmine, a 

 heavier fluid, is accomplished in a similar manner, but by means of 

 reservoir a'. This latter, held in a vertical position, is filled halfway to 

 the top of the short arm with carmine, and the whole of the carmine is 

 then sucked into the long arm, which is now hermetically closed at the 

 top, either with the finger or (better) a small stopper. The now empty 

 short arm is filled up as follows : Make three mixtures — 



No. 1 . Equal parts carmine and 30 per cent, alcohol . 50 per cent, carmine 

 No. 2. „ „ No. 1 „ „ „ . 25 „ „ 



No. 3. „ „ „ „ carmine ... 75 „ „ 



Fill the short arm up to 1 with No. 3, from 1 to 2 with No. 1, from 2 to 3 

 with No. 2, and from 3 to the top with 30 per cent, alcohol. Now move 

 the box and filter from a to a', avoiding bubbles, and finally remove the 

 hermetic seal from the long arm of a' and fill it up with carmine. The 

 flow begins as before, but the object is now transferred from the lighter 

 30 per cent, alcohol to the heavier borax-carmine, a change which would 

 be impracticable in Form A. 



The differentiator being now described, it remains to add some sugges- 

 tions brought to mind by my experience with it : 



1. Caoutchouc is soluble in turpentine, ether, &c. In case any of 

 these fluids are used the caoutchouc tubing should be fresh and thick, 

 and, further, be firmly tied to the glass parts. 



2. Small objects get lost in the cotton. If, however, each plug is 

 done up in fine, new linen cloth, no loss will occur. 



3. Notes, if plainly written with pencil, are legible even after passing 

 through carmine. 



4. In making a' the glass tubing must be gradually and thoroughly 

 heated in a good Bansen flame. The capillary termination of the filter is 



n 2 



