THE FLIGHT ACTIVITIES OF THE HONEYBEE 9 
some trouble in case the mercuric oxide collected as a flocculent mass 
on the platinum points. Under such circumstances the weight of the 
bee is no longer sufficient to overcome the greater displacement of 
mercury now necessary, and multiple contacts may result. This 
error was apparently eliminated by adjusting the apparatus so as 
to bring the mercury cups closer to the fulcrum, giving the bee a 
greater moment than this opposing force. 
5. The bees about to enter the tunnel may push against the closed 
inner door (especially if its clearance is large) and may occasionally 
produce multiple contacts when the tunnel is either on its downward 
or on its upward stroke. 
6. A bee pushing hard against the glass of an outgoing gate may 
produce multiple contacts by preventing a tunnel from falling 
rapidly past its critical point. Errors from this cause and from 
pushing against the closed inner door were very much reduced by 
darkening the glass on the gates (p. 8). 
7. Bees clawing at the closed outer door of the device may bring 
the tunnel down far enough to form an electrical contact, especially 
if the doors are roughened by dirt or propolis or by a corrosion of 
the surface of the metal. Error from this source was peculiar to the 
ingoing gates, for on only two occasions was it observed on an out- 
going gate. An attempt was made to eliminate this source of error 
by placing small metal cups over the exit hole so that clawing could 
take place only in an upward direction. 
8. The segments of some of the recording cyclometers would some- 
times bind against each other, thus failing to record the contacts 
when the magnet was excited by the passage of a bee. This necessi- 
tated some readjustments of the cyclometers. 
9. The cleaning of the gates in the evening, after the day's records 
had been taken, necessitated the use of an artificial light, which on 
warm nights attracted some bees from inside the hive. These might 
remain outside all night, thus introducing an error on the following day. 
10. The clustering out of the bees in the warm weather following 
the main honey flow was one of the most disturbing features of this 
investigation. On many days, records which were normal during 
the early part of the day were rendered valueless later on through 
the clustering out of the bees in the afternoon. At the same time 
this clustering out occurred throughout the apiary. 
11. In the construction of the outgoing gates a small strip of metal 
was left between the outer aperture and the lower edge of the glass 
window, which was sometimes grasped by the bee in its fall, giving 
rise to multiple contacts. 
12. A small error was caused by the actual drifting of the bees. 
This was apparent on colder mornings when the other hives in the 
apiary were active and the experimental one had not yet commenced 
its activity. 
It is clear from the preceding list of causes of errors that with the 
exception of No. 8, and sometimes of No. 1, all these factors have a 
tendency to increase rather than decrease the recorded exits and 
entrances of the bees. With the experience gained in the 1922 
season in the design and handling of this apparatus, the writer is 
convinced that, with the exception of Nos. 2 and 3, where debris is 
the cause of error, all these factors may eventually be eliminated. 
26909°— 25 2 
