Siete

(SHesieiiest

The temperature of the honeybee cluster

Contribution from the Bureau of Entomology, L. O. Howard, Chief.
* April 30, 1914.

THE TEMPERATURE OF THE HONEYBEE CLUSTER

— IN WINTER.
By E. F./ Puituies, Ph. D., In Charge of Bee Culture Investigations, and
Grorce 8S. DemuTtH, Apicultural Assistant.
INTRODUCTION.

The care of bees in winter is one of the most perplexing problems

confronting the beekeeper, especially in the North. This appears to
be due chiefly to the fact that it is difficult to determine by direct ob-
servation the normal activities of the bee colony in winter, and conse-
quently it is well-nigh impossible to determine what external condi-
tions are most favorable except by the gross results of experience.
Nor can we by a study of our wintering successes and failures deter-
mine definitely whether the same conditions of temperature and
humidity are desirable throughout the entire winter.’ On account,
therefore, of the lack of accurate knowledge of the activities of bees
in the winter season this problem has been taken up with the aid of
certain special apparatus and equipment. This preliminary report is
not to be considered as giving definite recommendations as to the care
of bees in winter, but rather.is issued to make known to Beekeeyes
some of the interesting results obtained in the first season’s work on
the behavior of the bees during the winter season.
' American beekeepers lose thousands of dollars annually in winter
from the actual death of colonies and even still more from those
colonies that do not die, but which are reduced in numbers and
vitality. The wintering problem is therefore a vitally important one.
The factors influencing the welfare of the colony and the behavior
lof the bees are numerous and closely interrelated. Of the chief ones
may be mentioned external temperature, food, ventilation, humidity,
the condition of the colony at the beginning of winter, and various
forms of irritation. In the present paper special emphasis is placed
on heat production, by which is meant the responses of the bees of
the cluster to the outer temperature and to changes in the outer tem-
perature as manifested in the generation of heat by the bees.

Norn.—This bulletin presents studies of bees as affected by temperature conditions
during winter and is of special interest to beekeepers inthe North.

36157°—Bull. 93—14 n

*

3 v of oan ey vi
2 BULLETIN 93, vo Spnekleiteyte/or 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 almos
impossible to avoid disturbing the cluster at each condi so that i
reacts abnormally. Furthermore, as the authors will attempt t
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 pV 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 1919-13, 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° B, far more uniform than the ordinary
bee cellar. Abundant wontilation 3 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,

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 3

only after the day’s records are made) and only when absolutely
necessary. It is found that entering the constant-temperature room
may under some conditions influence the behavior of the bees in a
marked manner.

Other colonies are kept on the roof of the same laboratory, where
they are left untouched from the beginning to the end of a series
of readings. The wires of the thermometers are led to the room
below through rubber tubes, and all the temperature readings are
made at a distance, as is absolutely necessary to eliminate disturbance.
Disturbances of outside colonies have also been found to influence
their behavior in a pronounced manner, especially in cold weather.

By studying the temperature of various fixed points within each
hive it has been found possible to use the temperature readings as a
substitute for direct observations. After becoming familiar with
the normal temperature and the temperatures incident to various
activities one can. ‘tell the shape, location, and various activities of
the cluster by a study of the temperature of different points within
the hive and can, in fact, form an opinion as to the welfare of the
colony. It has therefore been possible to follow closely the activities
of each cluster without opening the hives and even without going
near them.

THE INFLUENCE OF EXTERNAL TEMPERATURE ON HEAT
PRODUCTION.

The colony (A) to be discussed under this heading was wintered
out of doors (1912-13) on the roof, where the bees were free to fly
whenever the weather permitted. It was in a 10-frame Langstroth
hive, with the entrance reduced to 3 inch deep and 8 inches wide, and
was not packed or given additional protection. This hive contained
19 of the electrical thermometers—12 among the combs, 4 in the cor-
ners of the hive, and 3 on the bottom board. Readings were made
hourly from 9 a. m. to 4 p. m. through the winter (Sept. 26 to Mar.
28), except Sundays and holidays, and at intervals additional read-
ings were made every 15 minutes (or sometimes every 30 minutes)
during the night (5 p. m. to 8.45 a. m.) for periods of several days
each. In all, 41,413 temperature records were made for colony A.

The reaction of the cluster in heat production, as induced by
changes in external temperature, is well shown by the records made
from noon November 13 to 2 p. m., November 15 (1912), when read-
ings were made hourly from 9 a. m. to 4 p. m. and every 15 minutes
at night. From noon on November 13 the outside temperature
dropped slowly until 6 a. m., November 15, and the weather was
cloudy, so that the bees did not fly. It will be seen from the accom-
panying diagram (fig. 1) that at noon on the 13th the outside tem-
‘perature was about 69.2° F. and all the points within the hive were

BULLETIN 93, U. S. DEPARTMENT OF AGRICULTURE.

LDECKELS LAPTRENATETT-

‘ ‘udT}ONpord zoey 4Sag JO OUT} oq} 7B Je}SNID 94} JO 98pa 393no0
aq} TO safdnod Jo seinjVIadmMsa, AMOYS SOUT] JIOYg aut] sABoy 94} AQ payBOIPUL e1B sSedNjeaedmI9} JOOPINO “ZIGT ‘GE “AON
‘ard % 01 EL “AON UoOM MOIZ YW AUOTOD UT s90q JO Jajsn[o eq} JO 19}U99 94} JO a1n}VIeduI9} PUB BIN} BIeduIeZ 10OPjnNo 9G L—'T "SW

AIFSHNSANS STFAHOTO

oo) j oS
a fo) NK =
is 7 we
a liner -
OF N LOS
Aa boo
eS am
os 9S
oS DA a
ee aN < pel
_— tN Te i =
a aL = ed = we
- a ye Vi = lee
we Ze

/ —— ie
OL Oe
1 | = 74
a 7 E if Le
8 t vg
oY .
wie AG ih | ie
ae \y | ee
ve \ io
082 iT 89
06 06
id Fa é Re é 7 5 o & Ea e g é 7 e $ e ee

| We > We | We We
G/ AION 2/ AION €7O/ EY AION

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 5

then cooler than the outside air, due to the fact that it took some
time for the inside of the hive to warm up. At 4 p. m. the outside
temperature had dropped to 65.3° F., when it was lower than any
of the points within the cluster, which had in the meantime become
warmer. From this time until 6 p. m. the next day (14th) the tem-
perature within the cluster gradually dropped as the outer air cooled,
until the lowest one (No. 9) was 57° F. (Outside temperature,
48.2° F.) The generation of heat began at 6.15 p. m. at this point,
which was to one side of the cluster, and is to be attributed to the
movement of the bees in forming a definite cluster. At 6.30 p. m.
a rise in temperature was noticed on thermometer 19, at the other
side of the cluster. Until 10.15 p. m. the changes in temperature
are probably to be interpreted as incidental to the formation of a
compact cluster, and from this time until the next day at the close
of the series of readings the thermometers within the cluster showed
a considerably higher temperature than the outer dir, or than the
thermometers outside the cluster. The maximum in this series
was reached at 3.15 a. m., November 15, when thermometer 12 in
the center of the cluster registered over 89.4° F.

After the coldest outside temperature was reached and the outer
air began to get warmer (6.15 a. m., November 15), there was a
tendency for the cluster temperatures to drop. This is somewhat
noticeable in the case now being discussed, and is more clearly seen
in records obtained in other series. In general, after a period of
cold, when the outside temperature begins to rise, the cluster temper-
atures drop slowly to meet the outside temperature. The generation
of heat is reduced, or even discontinued, only to be increased when
the outside temperature again drops, or when it gets high enough to
induce greater activity, as in flight. It is found also by taking more
frequent readings when the cluster temperature is above about 69° F.
that it is less constant than when it is below this temperature, indi-
cating that at temperatures above this point the bees move about to
some extent, while between 57° and 69° they are quiet, unless flight i is
desirable owing to a long confinement.

This series of readings is supported by numerous records taken on
this and other colonies throughout the winter and, since all the ob-
servations tend to confirm what was first seen on the record pre-
sented here, the authors feel justified in presenting a definite state-
ment of the reactions of the cluster to outside temperatures. -It may
be added that a careful study of the records of previous investigators
fails to show a similar statement on this subject. When a colony is
without brood, if the bees do not fly and are not disturbed and if the
temperature does not go too high, the bees generate practically no
heat until the coolest point among the bees reaches a temperature of
about 57° F. At temperatures above 57° F. a compact cluster is
not formed, but the bees are widely distributed over the combs. At

6 BULLETIN 93, U. §. DEPARTMENT OF AGRICULTURE.

the lower critical temperature, which is for the present stated as
57° F., the bees begin to form a compact cluster, and if the tempera-
ture of the air surrounding them continues to drop they begin to gen-
erate heat within the cluster, often reaching temperatures consider-
ably higher than those at which they were formerly quiet and satis-
_ fied. It is evident, therefore, that the temperature within the cluster
is far from being uniform in winter, as has been, in a sense, assumed
among practical beekeepers. At the temperature at which other in-
sects become less active (begin hibernation) the honeybee becomes
more active and generates heat, in some cases until the temperature
within the cluster is as high as that of the brood nest in summer. To_
sum up, when the temperature of a colony of undisturbed broodless
bees is above 57° F. and below about 69° F. the bees are quiet and
their temperature drifts with the outer temperature; at lower tem-
peratures they form a compact cluster, and the temperature ‘within it
is raised by heat generated by the bees.

The authors desire to state that while the lower critical point,
57° F., appears rather well established, the observations up to the
present do not justify too definite a statement concerning the upper
limit of quiescence. It must be emphasized that these conditions do
not apply when the colony has brood. The rearing of brood in win-
ter causes a marked increase in heat production and constitutes a
condition which may becomé one of the most disastrous that can be-
fall a confined colony. This will be discussed at a later time.

When the heat production of the colony is explained, we are able
to understand to some extent the divergence in the records obtained
by other observers. It has, of course, long been known that bees
generate heat, and it has been pointed out that during cold weather
the temperature of the cluster is often higher than during warmer
weather. While the temperatures previously recorded are in most
cases abnormal, due to disturbance, the chief difficulty in under-
standing the phenomena which take place is due to insufficient ob-
servations. For example, if between noon November 13 and 2 p. m.
November 15 only a half dozen temperature records had been made
for the cluster (and perhaps without finding the warmest part of it)
and the outside air, it would have been impossible to determine the
limits of heat production. Most observers have been satisfied with
a few observations, and seemingly everyone who has inserted a ther-
mometer in a hive has felt called upon to publish the results, thereby
only confusing the problem.

THE EFFECT OF CONFINEMENT AND THE ACCUMULATION OF
FECES.

Before beginning a discussion of the effect of confinement and the
accumulation of feces, it may be recalled that during the active
summer season the length of life of worker bees is in a sense deter-

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. a

mined by the work done by them, rather than by days or weeks.
The greater the necessity for excessive activity the shorter the term
of life. The authors believe that they have evidence to prove that
this applies to the winter also, and this belief is entirely supported
by the experience of beekeepers everywhere. That bees may come
out of winter quarters strong in numbers and vitality it follows that
the work to be done by the bees in the winter should be reduced to a
minimum; and the winter problem, as thus interpreted, is thereforé
to find the conditions under which broodless bees do the least work.
The work which broodless bees do in winter consists, so far as has
been determined, solely in the production of heat or in activity inci-
dent to flying on warm days (if free to fly), and therefore the prob-
lem, so far as it is under the control of the beekeeper, is primarily
to obviate the necessity for the production of heat. If brood is
reared the work of the bees is necessarily enormously increased, and
their vitality is correspondingly decreased. So far as evidence is
available in this work, the colony is not fully recompensed for this
expenditure of energy by an increase in the strength of the colony
by bees thus reared,

The colonies: to be discussed under this heading (Nos. 1 and 3)
were wintered in the constant-temperature room in special 6-frame
hives (to economize space and concentrate the colony so that fewer
thermometers would be required) with full entrances and were not
propolized or sealed at the top. During the regular series of read-
ings the room was kept at a temperature which rarely dropped below
40° F. or went above 45° F., and the average temperature from
October 14 to March 6 was 42.67° F. This temperature was chosen
as being nearly the one usually considered best by beekeepers. The
foods given these colonies were stored in the combs, just as placed
by the bees. There was some pollen available in colony No. 1.
(Fig. 2.)

According to what has been said in the previous section, we should
expect bees at such a temperature to maintain a compact cluster and
to generate some heat at all times. This was actually the case, the
temperature of the interior of the clusters dropping below 64° F.
only a few times in either colony.

Colony No. 1, on honey stores, was in the constant-temperature
room from- Ostober 12, 1912, to March 24, 1913, or 168 days.? It

1In order that the young bees might all get a flight before the winter confinement, the
two colonies here discussed were placed in the constant-temperature room after the brood
had been removed. They were kept here several days, removed for a flight, and then
returned to the room for the regular series. The significance of this manipulation must
be reserved for a later discussion. This explanation is made to show how it was possible
to put these colonies in the room so early in a climate as mild as that at Philadelphia.
The object was, of course, to increase the time available for observation. Bees are
usually not wintered in cellars in climates ag mild as that of Philadelphia.

2In all, 24,077 temperature records were made for this colony.

8 BULLETIN 93, U. S. DEPARTMENT OF AGRICULTURE.

was then removed for a flight and put back the same evening, where
it remained until March 28. From March 7 at 9 a. m. until March
28 at 4 p. m. readings were made on this colony every 15 minutes
night and day, with the exception of the period between 9 a. m.
and 7 p. m. on the 24th, when it was out of doors. During this
period of three weeks the temperature of the room was changed:
slowly, being raised as high as 64° F. and cooled to 13° F.

a 19 2 L he ae a 7 ae 2a | vee | os ! 6
a lala i ae
oS" g \ 88"
oo" ‘a zi i { e6*
aes y I AI bss
e i i TN ad Te
e il i TAY TN To
oe | ly I \ iY TT Mt felis
Kala A Tala TV ms
ee BL: AY PT el :
7 yh ] y an
S + \ A zo
N oF A N\A u I 6a° a
SoA ALL ATT an
ye a a | ee
pc | =
We me
So" |__| ses Ny
qe aN
50" [ a
at DAL AT ia a Val F=RI AINE RT aes
«oil V f if vi a=

Fic, 2.—Average daily temperatures of the center of the cluster of bees in colonies 1
and 3 and room temperatures, Oct. 14, 1912, to Mar. 6, 1913. Taken from readings made
hourly from 9 a. m, to 4 p, m. The room temperatures are indicated by the heavy line.

When this colony was first placed in the room for the regular
series of readings, after a preliminary confinement, October 12 (the
readings were begun Monday, Oct. 14), it maintained a cluster tem-
perature which usually lay between 64° and 68° F., the daily average
temperature departing from these rather narrow limits only four times
up te November 22. The average temperature is 66.5° F. During the
first five weeks the temperature of the room was less regular than
later (due to faulty working of the regulating apparatus), and this

doubtless accounts for some irregularities in the cluster temperature. °° °

At first the three thermometers in the cluster (1, 2, and 5) gave tem-

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 9

perature readings quite close together, while thermometer 6, which
was near the cluster, gave readings intermediate between the three
thermometers of the cluster and the four others in the hive, farther
from the cluster. After November 22 the records of the thermometers
in the cluster were more widely separated and the temperature of the
center of the cluster (shown on thermometer 5) tended to rise gradu-
ally. It varied constantly, but by December 7 and from then until
the end of the month, it averaged between 69° and 75° F. On Novem-
ber 29 and December 12 the cluster temperature rose to over 88° F.
From the ist of January until March 6, which ended the regular
series of readings, the cluster temperature became more and more
irregular, and on January 20 the cluster moved (probably to accom-
modate itself to the stores) until thermometer 2 was nearer the center
and showed a higher temperature than thermometer 5. The size of
the cluster was gradually decreased by the death of bees, and all the
thermometers except 2 and 6 show a gradual decrease in temperature
until finally, from about February 25 to March 6, they are all low
and of nearly equal temperature. The two thermometers giving high
readings continued to show in general a higher and higher average
temperature and to become more irregular (except from February 15
to March 1), the periods of increased heat becoming more frequent.
There was absolutely no regularity in these intervals, After Febru-
ary 1 the temperature of the cluster varied between 75° and 91° F.,
the average from February 1 being 85.4° F.

On March 6 all colonies in the constant-temperature room except
two were removed. The tolony described above (No. 1) and one
other (No. 12), not to be described at present, were left. On March
7 at 9 a. m. the temperature of the room stood at 42° F., and the
temperature of the interior of the cluster was about 84° F. The
brine which cooled the room was then shut off and the temperature
of the room rose very slowly and regularly, until on March 11 at
8.45 a. m. it was 64° F. For the first day the temperature of the
cluster was slightly variable, and at 10.45 p. m. thermometer 6, which
had been cooler than thermometer 2, showed a rise in temperature
(probably due to a shifting of the cluster), and from then on to the
24th they were nearly of the same temperature at all times. On
March 8, at 3 a. m., thermometer 2 rose to 87° F. (room temperature,
48.5° F. ), having previously shown a cooling. The cluster tempera-
ture then dropped slightly, showing relatively little variation until
at 4.15 p. m., March 9, it stood at 77.3° F. (room temperature,
55.7° F.). As the room temperature continued to rise, the cluster
temperature increased still more rapidly, until at 8.15 a. m., March
11, it reached 93° F. (room temperature, 64.2° F.). A little brine
was now turned on, sufficient to lower the temperature gradually to
58° F. at 9 a. m., March 12, and it again rose to 63.8° F. at 5.45 p. m.,

10 BULLETIN 93, U. S. DEPARTMENT OF AGRICULTURE.

March 15. During this period the cluster temperature followed the
room temperature, but remained constantly over 20° warmer. The
room was again cooled slowly, and the cluster temperature dropped
until on March 16, at 3 p. m., the room was 49° F. and the cluster
77.5° F. As the room continued to cool, the cluster temperature 1n-
creased, the bees responding to the colder temperature, until at 4.15
a. m., March 17, the room was 48° F. and the ¢luster 88° F. The
room then gradually warmed, and again the temperature of the
cluster dropped and then again rose with the room temperature,
remaining always over 20° warmer. At 6.45 p. m., March 19, the
brine was turned on full and the room cooled rapidly, reaching the
minimum of 13° F. at 9 p. m., March 20. At no time, however,
did any of the thermometers in the hive record a temperature below
33° F. Here it remained constant within 0.1° F. for about six
hours, during which time the cluster temperature varied between
86.5° and 89.5° F. (a difference between the room and the cluster
temperatures of 73° to 76° F.). The brine was now shut off and
the room again warmed until 9 a. m., March 24, when it reached a
temperature of 44.5° F. During this warming the cluster cooled
until at the close it was varying between 72° and 79° F.

As stated above, the colony was now (9 a. m., March 24) removed
for a flight and put back the same day at 7 p.m. In the meantime
the room was cooled to 83° F. When the bees were put back into the
room the temperature of the entire inside of the hive showed great
variation and naturally an increase due to the warming up while out
of doors and to the activities of a good flight. The points outside
the cluster dropped rapidly, but it was midnight, March 25 (31
hours), before the curves of temperature again appeared normal.
The room was slowly warmed to 63.2° F. at 6.30 p. m., March 26,
and then slightly cooled to 54° F. at 6 a. m., March 27, and again
warmed to 58.5° F. at the close of the series, 4 p. m., March 28.
After the flight the temperature of the cluster never dropped below
89.5° F., and the highest temperature reached was over 95° F. (soon
after the flight). Thermometer 6 remained high, but thermometer
2, which had previously been high, now approached the other ther-
mometers, probably due to a rapid loss of bees and to a decrease in
the number of bees during the flight. It must be recalled that these
bees had been confined for an abnormally long time and were sub-
jected to treatment which is at least unusual. After this colony was
taken from the room for the last time it was found that thermometer
6 was over a patch of larve, and, estimating as accurately as possible,
the eggs from which these hatched must have been laid at the time
when the room was coldest (March 20-21) and when the cluster tem-
perature was at its highest point. There had been no brood previ-
ously, according to the temperature records as compared with those

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 11

of this colony earlier and with those of other colonies, nor was there
much evidence of increased heat production due to the presence of
brood until after the flight. Probably no extra heat was produced
for the eggs, and possibly the hatching of the eggs was somewhat
delayed by the low outer temperature. The effects on the cluster
temperature which might be expected from a flight, in relieving the
accumulation of feces, were not observed, because brood rearing had
been begun.

Colony No. 3 was placed in the constant-temperature room October
12, 1912, after a good flight, and readings were begun cn Monday,
the 14th. In all, 2,165 temperature records were made on Colony 3.
The stores provided this colony consisted of honeydew honey, which
was gathered in the department apiary and which, since it granulated
almost at once, had been removed by melting up the combs which
contained it. After this operation it remained liquid. During the
summer of 1912 some of this honeydew honey was fed to a colony
in the open, during a dearth of nectar, and was stored in new combs
above the brood chamber, in which no cells of pollen were to be
found. After the second storing the honeydew honey was clear, well
ripened, and did not granulate. This colony was also in a 6-frame
hive, as previously described, and contained five thermometers (Nos.
14-18) among the combs. It is of course well known to beekeepers
that honeydew honey is not a good food for winter.

When this colony was first put into the constant-temperature room
it behaved much as did Colony No. 1, except that the temperature
varied between 69° and 78.7° F. for the first week, being slightly
higher and more variable than that of Colony No. 1. The second
week it remained much the same, the temperature, however, varying
between 69° and 80° F. From this time on the temperature of the
center of the cluster rose rapidly, never dropping below 79° F. from
October 29 almost to the close of the readings. After November 4
the temperature remained above 86° F., and after November 11 it
dropped below 89° F. only twice until the end. Thermometer 17
at first read about 4° below thermometer 14, but after November 11
they were, close together until November 25, when thermometer 17
began to cool rapidly, due to loss of bees, and after November 30
thermometer 14 cooled rapidly until, on December 9, it showed that
no more bees remained alive. From December 2 to 7, inclusive, there
was little heat generated, due to the scarcity of bees. It is of interest’
to observe the records of thermometer 16, near the cluster, but usually
outside of it. It at first showed a temperature but little higher than
the two thermometers away from the cluster, but on October 31 it
began to rise until, on November 12, it reached 80.5° F., when it was
doubtless covered by the bees. Even the two thermometers (15 and
18) clear to the back of the hive rose until, on November 13, they

12 BULLETIN 93, U. 8. DEPARTMENT OF AGRICULTURE.

recorded 61.5° F. These thermometers showed about the same tem-
peratures for about 10 days, and then these two and thermometer 16
showed a cooling, since the bees were dying so fast that there were
no longer enough to warm up these thermometers away from the
center of activity. It was to be expected that this colony would die,
and the experiment was performed to learn the phenomena incident «
to the loss.

Before summing up the results of these two colonies, Nos. 1 and 3,
it may be stated that, so far as the evidence here presented is con-
cerned, the results as far as here discussed are confirmed by records
from 10 other colonies kept. in the constant-temperature room, but
fed other foods and otherwise different. There is in all of the records
no evidence which the authors can interpret as at all contrary to the
views here stated. A discussion of these other colonies is reserved.

It is evident from the behavior of colony No. 1 that at least one
factor entered which gradually caused the bees in the cluster to
generate more and more heat until at the beginning of the special
series, March 7, the cluster temperature was about 20° warmer than
it was at the same room temperature at the beginning of the confine-
ment. It is also seen that during the special series, March 7-24, the
cluster temperature always remained at least 20° above the room
temperature, whereas from the discussion of bees unconfined
(Colony A) we might expect them to cease heat generation when
above the lower critical temperature (57° F.). In the case of colony
8, fed on honeydew honey stores, the factor which caused more heat
to be produced evidently increased much more rapidly. As stated
previously, honeydew honey is a poor food for winter and is so
recognized. It contains the same sugars as honey, but contains in
addition a considerable amount of dextrin, the particular lot fed. to
colony 3 containing 4.55 per cent while good honeys contain only a
fraction of 1 per cent. From the evidence at hand it appears that
dextrin can not be digested by bees and, whether or not this is the
explanation, honeydew honey causes a rapid accumulation of feces
which usually results in the condition known as dysentery, in bad
cases of which the feces are voided in the hive. In the case of colony
3 the whole hive inside and out, as well as the frames and combs, were
spotted badly, the inside of the hive being practically covered. Even
with fine honey stores such a spotting i is usually noticed after a pro-
longed confinement, especially in severe weather (or during brood
rearing). It therefore appears that the accumulation of feces acts
as an irritant, causing the bees to become more active and conse-
quently (see later section ) to maintain a higher temperature. We
are therefore justified in believing that the cause of poor wintering
on honeydew honey is due to excessive activity, resulting in the bees
wearing themselves out and ultimately in the death of the colony.

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 13

In the case of colonies on good stores (e. g., colony 1) the feces
accumulate more slowly and the excess activity is not so marked and
is induced more gradually. The accumulation of feces due to con-
finement causes increased activity and this in turn is the cause of
excessive heat production, resulting in a reduction in the vitality of
the bees.

It therefore follows that excessive activity causes the consumption
of more food, resulting in turn in more feces, so that colonies on
poor stores are traveling in a vicious circle, which, if the feces can
not be discharged, results in the death of the colony. In the work
here recorded no attention was paid to the theory that dysentery is
due to an infection, since there is nothing in the observations made
that lends any support to that idea. If there is more than one kind
of dysentery, as has been held, then the observations here recorded
must be considered as applying only to the type which can be in-
duced at will in any confined colony by giving poor food and which,
as has been long recognized, can be relieved at once by an opportu-
nity for flight.

While the activity of the cluster is greater at some times than at
others, there are not, as has been held, regular intervals of activity
at which the colony rouses itself to take food. At no time is a colony
kept at a room temperature of 45° F. or less in a condition which
can be characterized as inactive. Presumably the reported “ inter-
vals of activity” have occurred when the colony made a noise due
to disturbance by the beekeeper.

The bees in colony 3 were compelled to, work constantly to main-
tain so high a cluster temperature. In fact, they did more work
than colonies wintered in the open air. Keeping these bees in a cel-
lar protected them from low outside temperatures, but the lack of
opportunity for a normal ejection of feces caused a condition more
serious than extreme cold weather. We seem to have here an expla-
nation of the fact, often observed by beekeepers, that‘ some colonies
wintered in the cellar are in worse condition in the spring than col-
onies that are exposed to severe cold. Poor food is evidently a more
serious handicap than low temperature.

METHODS OF HEAT PRODUCTION AND CONSERVATION.

A colony of bees in cold weather forms a compact, approximately
spherical ¢luster, but this cluster is not, as is usually believed, uni-
formly compact. In order to study the formation of the cluster and
as an aid to interpreting the temperature records in terms of action,
a colony (C) was placed out of doors in a narrow hive with double-
glass sides and top, and the stores were so arranged that the only
space available for the formation of the cluster was next to the glass
on one side, where it could ‘be kept under direct ‘observation. Since

14 BULLETIN 93, U. 8. DEPARTMENT OF AGRICULTURE.

the bees did not have room for a spherical cluster, they formed a
ring on the glass. Thermometers were then placed close together in
the outside space, so that the temperatures of various points could
be determined as desired. This hive was on the roof, and, while one
person watched the bees, constant communication could be kept up
with the person reading the temperatures in the room below by means
of a telephone, arranged so that the hands of both observers were
free. This colony was of course in the light, but the normal cluster
was nevertheless observed. It was disturbed as little as possible.
The nearly spherical cluster of bees consists, between the combs
and sometimes above or below, them, of an outer shell of bees close
together with their heads toward the center. This ring may be several
layers thick. The position with the heads inward is typical, except
{ when condensed moisture drops on the cluster as it often does in cool
weather, when the bees at the top turn so that their heads are up-
ward. The bees in this outer shell are quiet except for an occasional
shifting of position: Inside this rather definite shell the bees be-
tween the combs are not so closé together nor are they headed in any
one way. Considerable movement, such as walking, moving the abdo-
men from side to side, and rapid fanning of the wings, takes place
inside the sphere and when a bee becomes unusually active the ad-
joining bees move away, leaving an open space in which it can move
freely. Two bees may often be seen tugging at each other. In
addition to the bees between the combs, placed as above described,
others are in the empty cells of the comb on which the cluster is
always formed, always with their heads in. A verification of these
statements is contained in the following observations, and the ex-
periment may easily be repeated by anyone. For the purpose of
obtaining a colony without combs for another experiment, a hive was
opened December 15, 1913, while the outside temperature was low
enough to cause the formation of a compact cluster. When the combs
were separated the circle of bees in the shell was clearly observed.
When a comb from the center of the cluster was shaken the active
= hee in the center of the circle dropped off readily, and those in the
outer shell which were somewhat sluggish were removed with more
difficulty. After this was done those occupying empty cells in the
center of the sphere backed out of the cells and were shaken off.
Finally those occupying cells in the border of the sphere backed out,
showing a well-marked circle on the combs. Evidently the bees in
the shell, whether in the cells or between the combs, are less active
than those in the interior of the cluster. Naturally such a manipu-
lation as this is not to be recommended, except for purposes of
demonstration.
It is clear from observations previously recorded that the highest
temperatures are those of points in the center of this shell, and this is

TEMPERATURE OF THE HONEYBEE CLUSTER IN WINTER. 15

to be expected, as the heat is generated here. The outer shell consti-
tutes an ideal insulator for the conservation of the heat, since the
bees arranged so close together form small dead air spaces in their
interlacing hairs, especially those of the thorax, and afford still more
insulation with their bodies. The abdomens of the bees in the outer
row are practically separate one from another, and must often be
exposed to severe cold. That this method of conserving heat is
effective is shown by observations on undisturbed colonies out of
- doors. For example, on January 14, 1914, there was at 9 a. m. a
difference of 68° F. between thermometers 14 (center of the sphere)
and 16 (outside the cluster) of Colony D, which were less than 4}
inches apart on the same level in the same space between combs, and
a difference of 75° F. between this couple and the bottom board 4}
inches below it. What this difference might sometimes be in colder
climates may be imagined. Examples of this kind might be multi-
plied indefinitely from the records of these experiments.

- The source of the heat of the cluster must, of course, be the oxi-
dation of the food consumed by the bees. The bee is classed as a
cold-blooded animal in that the temperature of the individual bee
is practically that of the surrounding medium. There is obviously,
from the records just given, no internal regulation of the tempera-
ture of the body such as is found in birds and mammals, for the
temperature of a broodless cluster varies greatly. From the obser-
vations made on the various colonies, especially Colony C, it is clear
that heat for the warming of the cluster is produced by muscular
activity. While, of course, some heat is doubtless liberated by other
life processes, this is practically negligible when bees are quiet, as
in Colony A when above 57° F. That higher temperatures may be
produced, greatly increased muscular activity is required, and in
Colony C in cold weather bees in the center of the shell of insulating
bees were seen fanning vigorously and executing other movements,
such as shaking and rapid respiration. We thus have the para-
doxical condition that bees fan to heat the cluster in winter as well
as to cool ,the hive in summer. Observations of this kind were
repeated beyond number, and this theory of the method of heat pro-
duction is entirely supported by the repeated observation of a hum-
ming noise from the cluster during cold weather.

A few details of the observations on Colony C may be of interest.
For example, one bee was observed fanning vigorously for 74 minutes
(9.58 to 10.004 a. m., Jan. 28) while the other bees kept a space cleared
for it. The temperature of the nearest thermometer rose 4° F. during
this time. At 9.52 this thermometer was almost a degree cooler than
at the time of greatest heat during the fanning. The rapidity of fan-
ning of the wings varied, and toward the end of the time it became
so slow that the outline of the wings was distinguishable. After the

16 BULLETIN 93, U. 8. DEPARTMENT OF AGRICULTURE.

excessive activity this bee stood in the same place for a time. Rapid
respiration may play a more important part in heat production than
at first appears. One bee was observed to breathe 21 times in 14
seconds and then cease the rapid respiration. On other occasions
50 or more bees would begin shaking their bodies from side to side.

Tt has been shown in earlier sections that feces in the rectum cause
irritation, which induces increased activity and causes greater heat
production. It has also been found that other kinds of irritation
bring about the same result, but a discussion of these points can not
be undertaken here. It is at least evident from the records obtained
in this work that colonies of bees in winter, either in cellars or out of
doors, should be disturbed as little as possible. This appears to apply
especially to cold weather out of doors or in the cellar, especially
after the colony has been confined for some time.

The facts mentioned concerning the ability of the bees to conserve
the heat generated will perhaps raise the question as to the tempera-
ture of the hive outside the cluster in cold weather, when the cluster
is compact. In the case of Colony A the temperature of the hive
outside the cluster was often practically as low as the outside tem-
perature. This colony was not packed and had a rather large en-
trance. If the cluster forms such an efficient insulator in itself it
might be presumed that packing about the hive is of little value and
that it might.even be harmful, in that it would not serve to conserve
heat and would prevent the heat from the sun from penetrating to
the cluster. This line of reasoning, however, does not follow, and
in any case it is unsafe to speculate about these things without more
facts. The effects of various forms of packing are being studied.

In closing it may be desirable again to state that too hasty con-
clusions must not be drawn from the facts here presented. For
example, the records on heat production might be interpreted as indi-
cating the desirability of a cellar temperature higher than beekeepers
usually believe to be best. Experiments to test such a theory are now
being carried on, and it is found that a broad statement as to the
best cellar temperature can not yet be.given. Under most conditions
colonies can not be brought to the critical temperature, 57° F., with-
out disturbance. It is hoped that more work will throw some much-
needed light on this important subject.

x

ADDITIONAL COPIES

OF THIS PUBLICATION MAY BE PROCURED FROM
THE SUPERINTENDENT OF DOCUMENTS
GOVERNMENT PRINTING OFFICE
WASHINGTON, D. C.

AT

5 CENTS PER COPY
Vv

WASHINGTON : GOVERNMENT PRINTING OFFICE : 1914