New Ways of Measuring the Incubation Period of Birds

NEW WAYS OF MEASURING THE INCUBATION PERIOD OF BIRDS

S. CHARLES KENDEIGH

As pointedout by Swanberg(1950) and Nice (1954) there has in the past been considerableconfusion concerningthe proper way to measure the length of the incubationperiod of birds. Of concernhere is not the entire length of time that the adult sits on a clutch of eggs,but the minimum lengthof time requiredto bring an average,newly laid eggto hatching under normal nest conditions. For practical purposesit is now gen- erally agreed that the incubationperiod shouldbe consideredthe time from the laying of the last egg to the hatching of the last young when all eggsin the clutch hatch. The confusionlies in the fact that the first eggsof a clutch may receive someheat and undergoa certain amount of developmentbefore the clutch is completedor full incubationbegins. The eggstend as a consequenceto hatch in the order laid, but the intervals betweenhatching of the eggsis shorter than the intervals between their laying so that the time that elapsesbetween laying of an egg and its hatchingprogressively decreases with eachadditional egg in the clutch. In thosespecies, relatively few in number,where full incubationbegins with the first egg laid, the intervals betweenhatching are more nearly equivalentto the intervals between their laying. Even when the adult bird is observedsitting on the eggsfrequently, or for long periodsof time, in the egg-layingperiod, this doesnot necessarily mean that the eggsare being subjectedto full incubation heat (Swanberg, 1950). This dependson the stage of developmentof the brood patch which is being formed at this time (Bailey, 1952). Activity recordson the adult bird and thermocouplesplaced in the nest show that in the House Wren, Troglodytesaedo.n, the specieshere under consideration,full incubationbehavior and temperatureare not fully establisheduntil about the time that the last egg of the clutch is laid (Kendeigh, 1952). The presentstudy is concernedwith the total heat applied to the eggs, as a measureof incubation time, and with the proportional amount of heat appliedduring the egg-laying,incubating, and hatchingperiods. The original data were securedbetween 1934 and 1936 at the Baldwin Bird ResearchLaboratory, near Cleveland, Ohio. The present analysis was supportedby a grant from the National ScienceFoundation.

PROCEDURE Thermocouples were inserted under the eggs in six nests of the House Wren; in four nests on the day the first egg was laid, in one nest on the day the second egg was laid, and in one nest on the day the third egg was laid. Each egg was numbered

453 The Auk, 80: 453-461. October, 1963 454 I•EI•DEIOlt,Measuring Incubation Period [ Vol.Auk 80 soon after it was laid. All except one nest contained six-egg clutches; the one excep- tion was a four-egg clutch. Continuous nest temperatures were recorded on a Leeds and Northrup recording potentiometer until all the eggs hatched. From these records, it was possible to determine not only the exact time required for the incubation of each egg, but also the egg temperature throughout the egg-laying, incubating, and hatching periods. The way this was done is fully explained elsewhere (Kendeigh, 1963) and need not be repeated here. LENGTH OF THE INCUBATION PERIOD The time of laying of the eggscould be determinedwithin a few minutes from the temperaturerecord. It appearedin all casesthat the eggs werelaid daily duringthe first attentiveperiod or visit to the nestin the morning,even thoughthe femaleordinarily spent the entire night in the box. The averagetime of egg-layingwas 0530 hours, with extremesof 0510 and 0600. The time of hatching was only occasionallydiscernible from the nest temperaturerecord, so that dependencehad to be placedon frequent nest observations.Hatching time was taken as midway in the period to the last time the unhatchedegg was observedexcept when night intervened. Hatching in the HouseWren appearsnever to occurat night. The time of hatching was ascertainedwithin a few minutes in several instances,but may have been as much as three hours off in other cases. Probably the averageaccuracy of determininghatching time was plus or minus one and one-half hours. Since each egg was numberedafter being laid, the order of hatching could be determined. Of the 34 eggsin the six nests, 30 hatched. Only one of these 30 eggsis known to have hatched out of order; egg number three in nest 10 hatched four hours before egg number two. There are a few instances,however, when more than one egg was found hatchedat the sameinspection of the nest. The incubationperiod, calculated from the time of laying to the time of hatching of the last egg in the clutch for five nests (the last egg did not hatch in one nest, no. 43A) is almost exactly 14 days, or more precisely 336 --- 3.5 hours (mean plus standard deviation). In another sample, of 35 clutches,the length of the incubationperiod also averages14 days. HATCHING PERIOD In six-eggclutches, egg two hatcheson the averageabout 6 hoursafter eggone; egg three, 15 hours; egg four, 24 hours; egg five, 32 hours; and egg six, 47 hours. The intervalsbetween hatching of successiveeggs are thus 6, 9, 9, 8, and 15 hours. The total time from laying to hatchingis: for egg one, 409 hours; egg two, 391 hours; egg three, 376 hours; egg four, 361 hours; egg five, 345 hours; and egg six, 336 hours. The shorter intervals between hatching of the first eggsin the clutch and the longer time that elapsesbetween their laying and hatchingcorn- Oct.1963 ] KENDEIGIt,Measuring Incubation Period 455 pared with the later eggsin the clutchare best explainedby the lower temperaturesto which they are subjectedduring the egg-layingperiod (Kendeigh, 1963). INCUBATION TEMPERATURE The meantemperature of the eggsduring the periodof full incubation averages34.9øC in the HouseWren, but varies slightly with temperature (Y----33.9-t-0.045X, where X is air temperature). In the six nests (numbers25, 49, 75, 43, 10, 43A) studied,the calculatedegg temperatures were 34.7ø, 34.7ø, 34.7ø, 34.9% 35.0ø, and 35.2øC respectively (Kendeigh, 1963). After the first eggshatch, the behaviorof the adult on the nest (now combiningincubation of the remainingeggs and broodingof the new young) much resemblesthat when only eggsare present. Egg tempera- turesremain about the same. During the egg-layingperiod, the temperature of the eggsis lower, but varies greatly, sinceit dependson the air temperature,the amountof insolationthe nest box receivesfrom the sun, and the often irregular attention the eggsreceive from the adult bird. Beforeone can evaluatethe significanceof different temperatureson the lengthof the incubationperiod, one mustknow how the rate of development of the embryovaries with temperature.

RATE OF DEVELOPMENT AND TEMPERATURE THRESHOLD The rate of embryonicdevelopment is statedto be proportionalto the rate of energymetabolism of the embryo (Barott, 1937), and this can be determined from the rate of oxygen absorption. The rate of oxygen uptake of HouseWren eggswas measuredat six stagesof incubation(0, 2, 4, 7, 10, 12 days) and at six temperatures(21.1 ø, 26.7ø, 32.2ø, 35.0ø, 37.8% 40.6øC) by meansof a constant-pressuremahomettic respirometer (Kendeigh,1940). Omitting the lessreliable data for freshlylaid eggs (0 days), the rate of oxygenabsorption at the otherstages of incubation varied linearly with temperatureaccording to the followingequations, where Y is the cubic cm of oxygenabsorbed per hour and X is the air temperature: 2 days, Y----O.OO225X-O.0370; 4 days, Y----0.0086X- 0.1489; 7 days, Y----O.O28X-0.537; 10 days, Y----0.05X-0.85; 12 days, Y ----O.07X - 1.14. The temperaturethresholds for developmentat the five ages,calculated by solvingthe aboveequations when Y = 0, are 16.4ø, 17.3ø, 19.2%17.0 ø, 16.3øCrespectively. The over-allaverage is 17.2øC. The embryois inac- tive at temperaturesbelow the thresholdbut is not necessarilykilled. Edwards (1902) measuredthe early growth of the embryoof the domesticfowl at varioustemperatures and placedthe temperaturethreshold for developmentof the embryoat 20ø to 21øC. Funk and Bie]lier (1944) 456 KENDEIGI-I,Measuring Incubation Period L[ Vol.Auk 80 criticized Edwards' work and indicated the thresholdto be much higher, 26.7øC. One could not expect full developmentto hatching at any of theselow temperatureseven thoughgrowth proceeds slowly. Sincethe eggtemperature during incubation in this domesticatedtropical species is 38øC or above,compared with 35øC in the HouseWren, it is not surpris- ing that the temperaturethreshold is alsohigher. Anotherindication that the domesticfowl is more sensitiveto a drop in temperaturethan the House Wren is that the rate of embryonicdevelopment decreases 8.2 per cent/øC from 39ø to 28øC (Edwards,1902), comparedto 4.8 per cent/ øC in the House Wren.

TOTAL HEAT REQUIREMENT When one knowsthe temperaturethreshold for developmentand that the speedor rate of developmentis proportionalto temperatureabove this thresholdover the entire rangeof temperaturesto which the eggsare nor- mally exposedin the nest,then one can calculatethe total heat applied to the eggs. Effective temperature (egg temperatureminus temperature threshold)multiplied by total hoursfor developmentgives number of degree-hours,a measurementof appliedheat. Since the last egg in clutchesis maintained at full incubation temperature until hatched,the total heat requirementis most accuratelycalcu- lated for this egg. In five clutches,all but no. 43A where the last egg did not hatch, the total degree-hoursfor the last egg averaged5,867 -+ 144 (Table 1). If this numberof degree-hoursof heat is requiredto hatchthe last egg in a clutch,one would expectit also would be requiredfor all other eggsin the clutch. The numberof degree-hoursthat accumulatefor each egg during the egg-layingperiod has been calculatedby subtractingthe temperature thresholdof 17.2øCfrom the averagetemperature of the eggduring each three-hourperiod day and night as measuredby the nest thermocouple (Kendeigh, 1963) and multiplying by the number of hours involved. These figures,given in Table 1, show clearly that the total degree-hours for hatchingis essentiallythe same for all eggsin the clutch. The average number of degree-hoursfor 18 eggs,other than the last ones in the six clutches,is 5,989 -+ 196. The differencebetween the mean numberof degree-hoursfor the last egg and for the other eggsof the clutch is not statistically significantby the t-test. There is a reciprocalor intercompensatingrelation between the number of degree-hoursthat accumulatebetween the threeportions of the nesting cycle. This is true both betweendifferent eggsin the sameclutch and in comparisonsbetween different clutches. There is, of course,no accumulation of heat for the last eggduring the egg-layingperiod and the amount Oct.1963 ] KENDEIGH,Measuring Incubation Period 457

TABLE 1 ACCUMULATION OF DEGREE-tIOURS OF HEAT AT VARIOUS STAGES IN TtIE NESTING CYCLE OF THE HOUSE WREN

Nest Egg Egg-laying Incubating Hatching Total number period• periode perioda

No. 25 4 195 5,782 137 6,114 5 191 5,782 155 6,128 6 0 5,782 258 6,040 No. 49 3 519 5,363 105 5,987 4 399 5,363 106 5,868 5 198 5,363 262 5,823 6 0 5,363 528 5,891 No. 75 2 632 5,004 97 5,733 3 566 5,004 141 5,711 4 411 5,004 449 5,864 5 248 5,004 461 5,713 6 0 5,004 703 5,707 No. 43 2 870 4,992 218 6,080 3 722 4,992 543 6,257 5 299 4,992 706 5,997 6 0 4,992 972 5,964 No. 10 2 745 5,079 249 6,073 3 609 5,079 1434 5,831 4 0 5,½79 860 5,939 No. 43A 2 1,266 5,085 107 6,458 3 1.016 5,085 107 6,208 4 730 5,085 128 5,943 5 386 5,085 534 6,005 6 0 5,085 949'• 6,0345

From hour egg laid to hour last egg in clutch laid. From hour last egg laid to hour first egg hatched. From hour first egg hatched to hour particular egg hatched. Egg number 3 hatched4 hours before egg number 2. Assumingthis egg had hatched at the end of 336 hours. that must accumulateduring the hatching period is therefore the greatest --for this reason it is the last to hatch, The earliest eggs laid, on the other hand, accumulatethe most heat during the egg-layingperiod and the least during the hatching period. The other eggs fall progressively along a gradient in between. On accountof unusuallylow air temperatures,egg temperaturesduring the laying period were lowestat nest 25 (average23.2øC). This is corre- lated with the least accumulationof heat during this time and the greatest during the incubatingperiod, The actual incubatingperiod at this nest was 13 days, comparedwith 12 days at nest 49, 11 days at nests 10, 43, 75, and only 10 daysat nest43A. Nest box 43A was fully exposedto the sun and to high air temperatures,and the averageegg temperaturedur- ing the laying periodwas 32.0øC. Egg number2 had accumulatednearly 458 KE•D•Oa,Measuring Incubation Period [ Vol.Auk 80

20 per cent of the necessarydegree-hours before egg number 6 was laid. In anotherpaper (Kendeigh, 1963), I showedthat in regard to House Wren eggsa degree-hourof heat is equivalentto 5.6 g-cal. This means that the eggmust absorb an additional5.6 g-calof heat per hour for each IøC rise in temperature.The total of 5,867 degree-hoursrequired to hatchan eggis equivalent,therefore, to 32,855g-cal (32.9 kcal). If the egg temperaturedrops below the thresholdof 17.2øC, a corresponding amount of heat is requiredto bring it back to the threshold,even though no embryonicgrowth is involved.

ENERGY REQUIREMENT FOR EMBRYONIC DEVELOPMENT The energyfor embryonicdevelopment does not comefrom the input of heat from the outsidebut from oxidationof the fats, carbohydrates, and, perhaps,proteins in the yolk and albumenof the eggitself. The rate and amount of energy that is releasedfor maintenanceand growth dependson the amountof oxygenabsorbed and this is readily measured. The rate of oxygenabsorption by the HouseWren embryois, as indicated above,directly proportional to the temperatureand the lengthof time that the egg is maintainedat that temperature.The degree-hour,therefore, indicatesthe mobilizationof a definiteamount of energyand a definite amountof embryonicgrowth. The total oxygenconsumption of the House Wren embryo from the time it is laid to hatching is calculatedas 184.3 cubic cm. This differs from the 149 cubiccm givenby Kendeigh(1940) wherethe calculation was erroneouslymade for a 13 insteadof a 14 day incubationperiod. With a respiratoryquotient of 0.72, this is equivalentto an energytrans- formationof 866 g-cal (0.9 kcal). The total energyutilization for completedevelopment to hatchinghas beenshown for the domesticfowl to be 20 kcal (Barott, 1937) and for the domestic(probably mallard) duck,39.7 kcal (Kashkin,1961). These are muchlarger species,the incubationperiods are much longer,and the younghatch in a precocialrather than an altricial state. It is important to note, however,that theseenergy outputs are constants.This is indi- catedby Barott when he states: "As the temperatureof incubationwas loweredthe rate of metabolismdecreased, but the time of incubationwas increasedso that the total, which is the productof thesetwo, remained constant." If 5,867 degree-hoursrepresents the total heat appliedto the eggof the HouseWren from laying to hatchingand the total energytransformation is 866 g-cal,then a singledegree-hour represents an energytransformation by the embryoof about0.148 g-cal. Similarlyif 32,855g-cal of heat from Oct.1963 ] KENDEma,Measuring Incubation Period 459 the outsideare requiredto enablethe embryoto transform866 g-cal of energyfor growth,the efficiencyratio is only 2.6 per cent.

DISCUSSION Requirementsfor the incubationof HouseWren eggsto hatchingin respectto time in the nest under normal conditionsis 14 days, or more accurately366 --- 3.$ hours; in respectto heat appliedit is $,$67- 144 degree-hoursor 32.9 kcal abovethe temperaturethreshold; and in respect to energy transformedit is 0.9 kcal. These measurementsare of increas- ing biologicalsignificance and importance. Before we can fully under- stand the thermodynamicsof incubation,the significanceof behavioral differencesbetween species, and the role that the environmentplays, it is very desirablethat measurementsof heat and energy requirementsfor incubationbe obtainedfor many species.This will requireconsiderable effort and refinementof technique. We have only approximatedcorrect measurementsof heat and energy requirementsin the House Wren. We need a better method of recording the eggtemperature throughout incubation than by puncturingthe shell and insertinga thermocouple,by usingartificial eggs,or simply by measuringthe nest temperature. Before effective temperaturecan be calculatedwe must know the temperaturethreshold. We have taken 17.2øC for the HouseWren, as indicated by our data on oxygenconsumption, but the thresholdfor actual embryonicgrowth may be different. The rate of embryonicdevelopment is not necessarilyrelated absolutely to the rate of oxygenabsorption, since oxygenmust be usedfor maintenanceas well as for developmentand the proportionbetween the two may vary at different temperatures.There shouldbe direct measurementof embryonicgrowth at varioustemperatures. At best, 17.2øC representsthe alpha threshold (extrapolationto zero rate of developmentof a regressionline obtainedat mediumtempera- tures) rather than the true threshold,which may be lower (Shelford, 1929). There may alsobe a decreasein rate of developmentat temperatures above some unknown maximum. When the percentagesof normal developmentthat take place in the domesticfowl at varioustemperatures (Edwards, 1902) are plotted, it is evidentthat the alpha thresholdis 26.7øC,but that somedevelopment takesplace down nearly to 20øC. It is of interestthat the alphathreshold thus determinedagrees exactly with the figure given by Funk and Biellier (1944) for "the minimum temperaturefor embryonicdevelopment." I calculateddegree-hours of heat (or actually "developmentalunits" as usedby Shelford,1929) for the HouseWren, using26.7øC as the temperature threshold,but allowing proportional amounts of developmentto 460 K•.m)•.•oa,Measuring Incubation Period [ Vol.Auk 80

20øC, and obtained total units of heat considerablyless than given above and of almost the same statistical reliability. However, incubation tem- peraturesin the domesticfowl and the relation of embryonicdevelopment to temperatureare quite different than in the House Wren, as indicated above,and I prefer not to basemy calculationson data availablefor the fowl. Finally, measurementsare desirable,for each different species,of the total energyrequirements for the growth of the embryoto hatching. This may be approximatedby measuringthe total oxygenabsorption of the egg, or may be obtained, as did Tangl (1903), by finding the difference betweenthe caloric value of the fresh egg and of the newly-hatchedchick, its membranes,and of the unusedyolk. No one as yet has divided the energy transformedduring the incubation period into that required for maintenanceor existenceand that for development.This is desirable. It is obviousthat ornithologistsare only at a thresholdof another sort in the kinds and amount of data that need to be obtained and in the opportunitiesfor detailed analysis of the forcesat work in growth, be- havior, and ecologicalrelationships.

SUMMARY The incubation period of the House Wren, measured in time units, is 14 days or 366 --- 3.5 hours. A definite sum total of heat above a temperature thresholdis required for hatchinga bird's egg. In the House Wren, this is estimatedto be 5,867 -+ 144 degree-hoursor 32.9 kcal. A definite sum total of energy transformationis required by the embryo to complete developmentto hatching. In the House Wren, the energyrequirements for the existenceand developmentof the embryo is estimatedat 0.9 kcal. These units representprogressively more preciseways of measuringthe incubation period in birds.

L•T•RATXJRE CITm)

B^•¾, R.E. 1952. The incubation patch of passerinebirds. Condor, 54: 121-136. B^ROT•, H. G. 1937. Effect of temperature, humidity, and other factors on hatch of hen's eggs and on energy metabolism of chick embryos. U.S. Dept. Agr., Tech. Bull. 553: 1--45. Enw•m)s, C.L. 1902. The physiologicalzero and the index of developmentfor the egg of the domesticfowl, Gallus domesticus.Amer. J. Physiol., 6-' 351-397. F•J•}c, E. M., and H. V. B•LL•a. 1944. The minimum temperature for embryonic developmentin the domesticfowl (Gallus domesticus).Poult. Sci., 23: 538-540. K^s•x•, V.V. 1961. Heat exchangeof bird eggs during incubation. Biophysics, 6(1): 57-63. Km•'•)•a, S.C. 1940. Factors affecting length of incubation. Auk, 57.' 499-513. K•)•c,•t, S.C. 1952. Parental care and its evolution in birds. Illinois Biol. Mon- ogr., 2•: i-x, 1-358. Oct.1963 ] KE•I)E/GI-I,Measuring Incubation Period 461

KENDEIGYI,S.C. 1963. Thermodynamics of incubation in the House Wren, TrogIo- dytes aedon. Proc. XIII Intern. Ornithol. Congr., 1963: 884-904. NtcE, M.M. 1954. Problemsof incubation periodsin North American birds. Con- dor, 56: 173-197. SHEL•ORD,V. E. 1929. Laboratory and field ecology. Baltimore, Williams & Wil- kins Co. SW^NBER0,P.O. 1950. On the conceptof "incubationperiods." V•r F•gelvirld, 9: 63-80. T^•GL, F. 1903. Beitrige zur Energetik der Ontogenese.I. Mittheilung. Die Ent- wicklungsarbeitim Vogelei. Arch. f. d. Ges. Physiol., 93: 327-376. Departmentof Zoology,University of Illinois, Champaign,Illinois.