July1996] ShortCommunications andCommentaries 705 emy of Natural Sciences, Philadelphia, and scurus).In The birds of North America, no. 15 (A. American Ornithologists' Union, Washington, Poole,P. Stettenheim,and F. Gill, Eds.).Academy D.C. of Natural Sciences,Philadelphia, and American ZI•t•tERMAN,J. 1963. A nesting study of the Catbird Ornithologists'Union, Washington,D.C. in southernMichigan. Jack-Pine Warbler 41:142- 160. Received10 July 1995,accepted 25 September1995. ZWICKEI,F. C. 1992. Blue Grouse(Dendragapus oh- The Auk 113(3):705-708, 1996 Influence of Lunar Cycle on Laying Dates of EuropeanNightjars ( Caprimulgus europaeus) C. M. PERRIlqS• AND H. Q. P. CRICl<2.3 EdwardGrey Institute of FieldOrnithology, Department of Zoology, SouthParks Road, Oxford OX1 3PS,United Kingdom; and BritishTrust for Ornithology,The National Centre for Ornithology, TheNunnery, Thetford, Norfolk IP24 2PU, UnitedKingdom Mills (1986) suggestedthat the timing of laying by another nocturnal caprimulgid, nests without rela- crepuscularor nocturnal birds, suchas caprimulgids, tion to the lunar cycle, which Brigham and Barclay might be related to the amount of ambient light with (! 992) suggestis related to their propensity to attempt different phasesof the moon. He argued that such two broods within a limited season. birds might be able to collectfood moreeasily at times In a brief report by Cresswell(1992), a relatively when the moon was near to full than when it was small sample(n = 30) of EuropeanNightjars in south- new. As part of a study of the Whip-poor-will (Ca- ern Englandwere shown to start egg laying around primulgusvociferus), he looked at a combined sample the time of the full moon in the first half of June, but of breeding recordsof five caprimulgid species,in- not if the full moon occurredin late May (and soalso cluding the European Nightjar (Caprimulguseuro- in late June).He suggestedthat there might be some paeus),and showed that there was indeed a nonran- advantagefor egg laying to occurduring a full moon dom distribution of laying dates.More clutcheswere in early June, but that to delay laying until a full started in the half-month prior to full moon than in moon in late June would limit the chances of suc- the other half. Furthermore, he found that Whip- cessfullyrearing a secondbrood. poor-wills foragedmore (seealso Brighamand Bar- The sample sizesin Cresswell'sand Mills' papers clay 1992), sang more, and fed their nestlings more were not very large (the latter had to combine data frequently on moonlit than on dark nights. Mills sug- from 79 recordsof five speciesof Caprimulgus,ranging gested that the timing of laying was synchronized from North America and Europeto South Africa). In with the lunar cycleso that the first two weeksof the this paper we examinethe relationship between lay- nestling period would coincide with the greatest ing and the lunar cycle for the EuropeanNightjar amount of moonlight. He also noted that this sched- using a much larger sample of recordsfrom Britain ule would result in high levels of moonlight five to (n = 464). six weeks after hatching when the young gain in- Methods.--Ouranalysis makes use of the Nest Rec- dependence from their parents. Similar results have ord Card data available from the British Trust for been found for the Fiery-neckedNightjar (C. pector- Ornithology. Nest Record Cards are completedby alis),which concentratesegg laying in the week fol- volunteer observers who record details of the nest lowing a full moon (Jackson1985). However, data for contentson each visit made to a nest (Mayer-Gross two speciesof nighthawk--the Common Nighthawk 1970). The information is coded and computerized (Chordeilesminor) and the LesserNighthawk (C. acu- and, for many records,laying date can be back-cal- tipennis)--showedno suchcorrelation between laying culatedon the basisof the nestcontents at later stages date and phase of lunar cycle, which Mills (1986) of nesting.The analysisis basedon 464 recordswith thought was becausethese speciesare partially di- sufficient data for us to estimate the date of clutch urnal. The CommonPoorwill (Phalaenoptilusnuttallii), initiation to an accuracyof + 5 days. Most (454, or 98%)of the recordscome from England,and the rest come from Wales; they were recorded from 1921 to Addresscorrespondence to this author. 1991. The English recordscome mainly from three 706 ShortCommunications andCommentaries [Auk,Vol. 113 25 A 20- 5- o o 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 E2q z 1- 0 0 2' 4' 6' 8' 1 '0 12' 14' 16' 18' 20' 2'2 24' 2'6 2'8 30 Lunor doy Fig. 1. Numbersof EuropeanNightjar nestsinitiated on eachday of the lunar month basedon data from (A) 464 nest recordsfrom all calendarmonths, and (B) 127 recordsfor which laying began between 5 and 18 June.Linear regression for the latter is: Y = 2.9 + 0.0919X,where Y is the numberof initiationsand X is the lunar day (F = 7.34, df = 1 and 28, P = 0.011). counties (Surrey, 119; Nottinghamshire, 102; and For each clutch, the date of the full moon prior to Hampshire,79). This representsa relatively restricted when the clutch was started was calculated. This was geographicarea stretching ca. 200 km north-south done by selectingthe date and time of a full moon over which environmental factors(such as length of and repeatedly subtracting the mean lunar month twilight) vary relatively little. from this so as to obtain the dates of the full moons July1996] ShortCommunications andCommentaries 707 throughout the period under study (Tennent 1990). ploratory analysisand is the main period of laying The difference between the date of the full moon of first clutchesin England (Nest RecordCard data). prior to the start of each clutch and the actual start A number of complicating factors might obscure of eachclutch provided the numberof "lunar days" sucha result at other times of the year. In a country delay betweenfull moonand the startof eachclutch. such as England where there often is cloud cover Quadraticand linear regressionswere usedto inves- during the early part of the laying seasonof European tigatethe relationshipbetween the numberof nests Nightjars,this might be moreimportant than the phase initiated and lunar date. Quadratic regressionwas of the moon in affectingtiming of laying. Although favoredif it provideda significantimprovement over we were not able to analyzethis factor,Mills (1986) linear regression.In addition, the median frequency found that cloud cover was not a significant factor of nestsrecorded on each lunar date was compared affectingactivity levels of Whip-poor-wills.Similarly, for the first (lunar days0-14) and second(lunar days other weather variables,such as rain, temperature, 15-29)halves of the lunar monthusing Wilcoxon two- and wind speed,all of which are likely to affectprey sample tests.Analyses were undertaken using SAS activity or visibility, might have effectson feeding programs(SAS Institute 1990). that couldobscure any generalrelationship between Results.--Therewas no significant linear or qua- laying date and moon phasethroughout the season. dratic relationshipbetween nest initiation and lunar Cresswelland Alexander (1992) showedthat the pro- date (Fig. 1A; P > 0.10), and there also was no sig- portion of time that EuropeanNightjars were active nificantdifference between median frequencies of nest in the middle of the night was inversely correlated initiations for the first and second halves of a lunar with the nightly minimum temperature.In addition, month (z = 1.45, P = 0.15). Birds losing their first the temperaturein the period prior to egg laying clutchesprobably cannot delay relaying until the might be important.Warm, cloudy weather might "right" time in relation to the next full moon, as too allow more insectsto emerge,whereas bright moonlit much potential breeding time might be lost. There- nights that are cool might have the oppositeeffect, fore, it may be more likely that only first clutches especiallyin an alreadycoolish climate. Mills (1986) would show any relationshipwith moon phase.For also noted that the duration of twilight after sunset clutchesinitiated in May, our resultswere similar to increaseswith latitude; on the equator it lastsfor 39 thosefor the whole data set (Fig. 1A), and there was min, while at 44ø N it lastsfor 84 min. In England, no sign of any relationshipbetween laying date and which is at higher latitudes (ca. 51ø N), the twilight phaseof moon. periodwill be longerstill, and the length of the night Finally, we testedto determineif Cresswell's(1992) in June is relatively short (i.e. 6-7 h). This might suggestionthat synchronizationbetween the moon explainthe lack of a relationshipbetween nest ini- and egg laying only occurredwhen the full moon tiation and phaseof the moon for clutchesstarted in fell in mid-June(5-18 June). We found a significant May becausenestling demands at thosenests would linear relationship(Fig. lB) of increasingfrequencies be highestin June when twilight and potential for- of nest initiations during the lunar month. Conse- aging time are longest.Brigham and Barclay(1992) quently, there were more nest initiationsper day in showedthat the foraging activity of CommonPoor- the second half of a lunar month than in the first (z wills increased with an increase in lunar light, but = 2.37, P = 0.018). Thus, in mid-June, fewer birds the nestingcycle was not synchronizedwith the lunar than expectedstart clutchesduring a waning moon, cycle.They suggestedthis was related to the relatively and more than expectedduring a waxing moon. high latitude of their study area (and, therefore, high Discussion.--TheEuropean Nightjar is a speciesthat availability of feeding time), in addition to their birds is thoughtto capturemost of its prey by silhouetting typically producingtwo broods. them againstthe light of the sky at night (Cramp Further evidencethat, for early clutches,European 1985).Thus, one might expectthat the easewith which Nightjars may not be dependent on the stage of the they cancatch their food (mainlyflying insects)would lunar cyclecomes from the work of Berry and Bibby be enhancedby lunar light.