12 THE VOYAGE OF H.M.S. CHALLENGER, 



bends of the stem, or forced into the bulb, when the instrument is exposed to very high 

 or very low temperatures. They seem to be in every respect objectionable, especially 

 as the necessity for them would be entirely removed by adding an inch or two to the 

 length of the instrument ; or, if they must be retained, by protecting them and using 

 more powerful magnets. Their presence produces an effect large compared with their 

 apparent importance. The sketch below represents, on a large scale, one of the most 

 highly developed of the more effective of these aneurisms, that which is situated close to 

 the main bulb of the instrument. 



By reason of the convexity of the thermometer tube the diameter of the bore appears 

 from the outside to be considerably larger than it really is. In fact a very simple 

 geometrical construction shows that the ratio of its apparent diameter to its real 

 diameter is that of the refractive index of glass to unity, i.e., it appears to be about 

 1'6 times its actual diameter. So that even when the aneurism, and the liquid filling 

 it, appear to occupy the whole diameter of the tube, they only occupy T ?g- or about two- 

 thirds, so that even in this extreme case the walls of the aneurism are not usually very 

 thin. The percentage diminution of volume of the middle portion of the aneurism is in 

 such a case (roughly) 50 per cent, greater than that of the unaltered tube. 



The real mischief done by the aneurism is not due mainly to thinness of the walls 

 and consequent greater liability to distortion by pressure ; it is due to the fact that the 

 aneurism, in consequence of its greater section, contains a much larger quantity of 

 mercury than does an equal length of the tube ; and therefore that a small percentage 

 diminution of its volume will produce a marked displacement by the outflow into the 

 narrow tube. Several of the aneurisms I have measured produce a disturbance of the 

 index corresponding to that produced by at least five times their own length of the tube. 



In some of the more exaggerated ones it actually produces an effect on the maximum 

 and minimum index equal to that due to the extension of very nearly one-half of the 

 mercury column in the thermometer. But this, though easily remediable, is not a defect 

 of much consequence. [The calculation of the effect due to an aneurism is given in 

 Appendix B.] 



Imploding and Exploding of the Thermometer Bulbs. 



In connection with the breaking of some of the thermometers, as a result of pressure 

 whether in the press or in the sea, it may be well to describe the curious nature of the 



