i 9 4 THE JOURNAL OF PHARMACOLOGY. 



It is not at all necessary to test a clinical thermometer throughout the whole scale 

 of degrees that is marked upon it. In most cases it will be sufficient to ascertain its accu- 

 racy at some point between about 97 and 104 F. But, if time is not pressing, it is best 

 to make at least tzvo tests, one at a temperature between about 97 and 102 F., and one 

 at a few degrees higher. If a thermometer agrees with the standard at both points 

 within one-fifth of a degree it may be regarded as sufficiently accurate for all practical 

 purposes. Any that agree with it exactly, or within one-tenth of a degree, may be 

 put aside as particularly accurate, and these may, of course, be sold at a higher price 

 with a corresponding certificate. 



Now, if only one clinical thermometer, or only a few, are to be tested, this may be 

 done simply by first shaking down the " index " in all of them and immersing them 

 with the standard — best tied together — in a rather large volume of water at the proper 

 temperature. If the thermometers are inserted cold, there will, of course, be a " cold 

 wave" produced around them, which will last until the temperature has become 

 equalized by diffusion or by stirring. If the volume of water was large, the final re- 

 sult will not vary by any appreciable degree from the theoretical one, because the 

 stored-up heat in the volume of water practically neutralized the slight chill intro- 

 duced by the few cold thermometers. 



In testing a large number of thermometers, however, at the same time — say one or 

 several gross,— another method must be adopted, one that will avoid, as much as pos- 

 sible, the production of the "cold wave" and the consequent chilling off of the water. 

 The following has been found by long experience to be the most simple and reliable 

 mode of procedure : 



First separate the thermometers which are to be examined into "high" and 

 "low" ones, according as the index or registering column stands high or low. All 

 " high " ones must be first manipulated so that the index is brought below the degree 

 at which the test is to be made. This may be done by shaking them either singly or 

 in bulk, in the latter case preferably enclosed in a bag or wadded box, or, more expe- 

 ditiously, by putting them in a centrifuge, such as is used for urinary or blood exam- 

 inations, and giving them a few rather gentle twirls, just sufficient to bring down the 

 index as far as required. Prolonged or violent twirling must be avoided, as it is apt 

 to ruin the "index." Fill two vessels of suitable capacity (see further on; with 

 water, one (A) at a temperature about one degree lower than that at which the ther- 

 mometers are to be tested, and the other (B) with water at the latter temperature. It 

 is not necessary to be very exact as to either temperature, as long as that in vessel A 

 is about a degree lower. The standard thermometer which accompanies the ther- 

 mometers to be tested will indicate what the exact temperature is. Place all the 

 thermometers, together with the standard, upon a wire frame, or other suitable sup- 

 port, permitting a free circulation of the water around them, and immerse the frame 

 into vessel A, which should contain a considerable volume of water, say about five 

 gallons when several gross of thermomete.s are to be tested. The volume in the ves- 

 sel B may be smaller. Allow the thermometers to remain in A during about five 

 minutes, which will be sufficient to bring them all up to the temperature of the 

 water. Meanwhile examine the temperature of the water in B, and, if necessary, 

 adjust it so that it is about one degree higher than in A, stirring well to insure homo- 

 geneity. Then raise the frame holding the thermometers out of vessel A and at once 

 transfer it to B standing close by, moving it up and down in the liquid a few times to 

 dispel any currents of unequal temperature. After a lapse of five minutes remove 

 the frame and examine the thermometers, if necessary, with a lens. They may now 

 be arranged in groups, according as they entirely agree with the standard, or differ 



