THE JOURNAL OF PHARMACOLOGY. 193 



hypercritical appliance, its annual production and consumption now reaches into the 

 millions. 



• Of course, the only practically useful clinical thermometer is the self-registering 

 one ; and this "may be of two kinds, either with plain front, or with magnifying front 

 lens. Of either of these there are again various grades or classes, differing as to the 

 shape of the bulb or stem, length of stem, color of glass, sensitiveness, etc., etc. It 

 is assumed that the reader is familiar with the several kinds ; hence a description is 

 deemed superfluous here. Moreover, for our purposes, all these distinctions can be 

 disregarded, since the method of testing their accuracy, which will be described 

 below, can be applied to all. 



As a rule.clinical thermometers are graduated between about 94 and no°F., in one- 

 fifth of a degree. For special purposes, where unusually high temperatures are to be 

 measured (as sometimes in cases of insolation, etc.), some may be had on which the 

 graduation is carried to H5°F. or even higher. 



If it is the object to examine the absolute accuracy of a thermometer for every 

 degree of its scale, the mode of procedure is much more complicated than that below 

 given, and requires special apparatus and considerable practical experience. For this 

 reason, so-called " standard " thermometers are rather expensive, and even the com- 

 mercial certificated clinical thermometers are rather high in price. Of course, if these 

 have been tested by a reliable authority, they are fully worth the money. But the 

 general public and most of the hospitals and charitable institutions in this country 

 cannot afford to pay this high price, and, therefore, have to content themselves with 

 uncertificated ones. But here they run the risk of getting worthless instruments. 

 While much trust may be placed in the names of certain manufacturers, it is never- 

 theless a fact that even the best brands will be found now and then — sometimes even 

 a good deal — deficient. 



Among the pharmacists in this country there are many who do a large business in 

 clinical thermometers, and whose professional reputation is of such a character that 

 their certification would be pccepted by medical men of their acquaintance, and 

 eventually also by others and the public, as a satisfactory guaranty of the accuracy of 

 any thermometers tested by themselves. There is, hence, a chance here of expanding 

 their business in a legitimate direction, and the writer hopes that this suggestion will 

 be followed by those who may find it worth their while. 



In the course of many years' experience the writer has examined many thousands 

 of thermometers, and has long ago settled down to a method which aims to avoid 

 errors of observation as far as possible, it is believed, and which permits to ascer- 

 tain the accuracy of instruments within limits that are considered sufficient in prac- 

 tice, namely, one-tenth of a degree, though for ordinary clinical purposes an accuracy 

 to one-fifth of a degree is amply sufficient. 



Of course, the first requisite for testing thermometers is the possession of a standard 

 thermometer. This need not be an expensive instrument, but may be any well- 

 seasoned clinical thermometer which has been carefully compared with a standard one 

 and found to be sufficiently accurate, or whose deviations from the standard are exactly 

 known for every degree of its sale. " Well-seasoned " in this case means that it must 

 have been made several years before it was graduated, as it is well known that the 

 calibre or bore of a thermometer suffers slight changes for some time after it has un- 

 dergone the process of manufacture. The best-reputed makers of thermometers 

 state that they never graduate and sell any thermometers that are not fully seasoned. 

 Unfortunately, there is no test by which a seasoned thermometer can be distinguished 

 from an useasoned one at the time of purchase. And while honest makers no doubt 

 adhere to the rule, it is well known that the majority disregard it altogether. 



