2 BULLETIN" 93, U. S. DEPARTMENT OF AGRICULTURE. 



A special reason for this emphasis in a preliminary paper is that 

 all previous work on the temperature of the cluster in winter, of 

 which there has been considerable, has failed to show definitely what 

 the normal responses are. The data are often those of abnormal con- 

 ditions and are therefore misleading, making them almost valueless 

 for purposes of application. One source of error which is to be 

 found in all the records known to the authors is the use of the mer- 

 cury thermometer, for, when such a thermometer is used, it is almost 

 impossible to avoid disturbing the cluster at each reading so that it 

 reacts abnormally. Furthermore, as the authors will attempt to 

 show at a later time, disturbances of the colony may influence the 

 temperature of the cluster for a considerable period, often more than 

 one day. Usually no account has been taken of the necessary cor- 

 rections to be made for the mercury thermometer. 



Because of the errors in other work on the subject, due to the use 

 of mercury thermometers, the thermometers chosen for the work 

 here recorded are of another kind. Electrical thermometers are used, 

 by means of which readings can be made without approaching the 

 hive, and the thermometers (couples) are of course permanently 

 fastened in place. These are of the type known as thermocouples 

 or thermal junctions and the readings are made by means of a poten- 

 tiometer indicator and a sensitive galvanometer of the d'Arsonval 

 type. The wires used in the thermocouples are copper and con- 

 stantin (a copper- nickel alloy) , giving an electromotive force of about 

 40 [i V per degree centigrade. A detailed description of the appara- 

 tus is impossible here, and it need only be stated that the method as 

 used gives readings to an accuracy of 0.09° F. (0.05° C.) ; the ther- 

 mometers are practically instantaneous in their action — that is, show 

 changes in temperature without a " lag " ; the readings of many 

 thermometers can be made consecutively on one carefully calibrated 

 instrument, insuring uniformity, which is impossible in using many 

 mercury thermometers; and, a point of importance in such work, 

 the readings can be made at the rate of two a minute, which would 

 be impossible with widely scattered instruments. In all, 161,617 tem- 

 perature readings were made during the winter 1912-13, and the 

 work is being continued. 



Part of the. colonies are kept in a well-insulated room (used as a 

 " bee cellar ") in the zoological laboratory of the University of Penn- 

 sylvania, Philadelphia, Pa., which can be kept at a temperature 

 usually varying not over 2° F., far more uniform than the ordinary 

 bee cellar. Abundant ventilation is provided, and the room is com- 

 pletely darkened to avoid possible disturbance by light. The tem- 

 peratures of the indoor colonies are read from an adjoining labora- 

 tory to eliminate the possible errors due to disturbance, and the room 

 is entered rarely (about once a week on an average and, if possible, 



