Hatching Dates for Common Sandpiper <I>Actitis Hypoleucos</I

Home , Actitis, Common sandpiper, Sandpiper

Hatching dates for Common Sandpiper Actitis hypoleucos chicks - variation with place and time

T.W. Dougall, P.K. Holland& D.W. Yalden

Dougall,T.W., Holland,P.K. & Yalden, D.W. 1995. Hatchingdates for CommonSandpiper Actitishypoleucos chicks - variationwith place and time. WaderStudy Group Bull. 76: 53-55. CommonSandpiper chicks hatched in 1990-94between 24 May (year-day146) and 13 July (year-day196), butthe averagehatch-date was variablebetween years, up to 10 days earlier in 1990 than in 1991. There are indicationsthat on average CommonSandpipers hatch a few days earlierin the Borders,the more northerlysite, butthis may reflecta changein the age structureof the Peak Districtpopulation between the 1970sand the 1980s- 1990s,perhaps the indirectconsequence of the bad weatherof April 1981. Dougall,T. W., 29 LaudstonGardens, Edinburgh EH3 9HJ, UK. Holland,P. K., 2 Rennie Court,Brettargh Drive, LancasterLA 1 5BN, UK. Yalden,D. W., Schoolof BiologicalSciences, University, Manchester M13 9PT, UK.

INTRODUCTION Yalden(1991a) in calculatingthe originalregression, and comingfrom the years 1977-1989 (mostlythe 1980s). CommonSandpipers Actitis hypoleucos have a short We also have, for comparison,the knownhatch dates for breedingseason, like mostwaders; arriving back from 49 nestsreported by Hollandet al. (1982), comingfrom West Africa in late April, most have laid eggs by mid-May, various sites in the Peak District in the 1970s. which hatcharound mid-June. Chicksfledge by early July, and by mid-Julymost breedingterritories are Ringingactivities continue through the breedingseason at deserted(Holland et al. 1982). The timingof the breeding both sites, and chickscan be at any age from 0 to 19 days season seems constantfrom year to year, but there are old when caught (thoughyoung chicks are generally few data to quantifythis impression.It is difficultto locate easier to find). Where severalchicks of a broodwere an adequatesample of nestsin any one year to provide caught,a singledate for hatchingwas entered. Older an adequateindication of the timingof the season,let chicks sometimessuggested hatchingdates two or three alone do so over several years. days apart, and the mean was usedfor the latter,the earlierdate (i.e. the apparenthatch date of the older The demonstrationby Holland& Yalden (1991) that bill chick)for the former. One mightexpect a broodof four lengths(and masses)of CommonSandpiper chicks are youngto take 1 - 1Y2days to hatch, so causing some stronglycorrelated with age offersan indirectapproach. inherentvariation and, of course,their growth rates may Reversingthe calculation,given bill lengthand the date of differ a little. ringing,the age and thereforethe hatch-dateof each chick couldbe estimated. Enoughchicks, at a range of ages The sampleof hatchdates is stronglyskewed in all throughthe breedingseason, can be caughteach season samples,so non-parametrictests (Meddis1984) have to allowcomparisons between sites and years. been usedto comparethe mean hatchdates for each sample(site or year). Hatch-datesare presentedas year- days, so that 1 June is year-day 152, or 153 in a leap MATERIALS AND METHODS year.

Holland& Yalden (1991a) showedthat bill length,y (in mm) was relatedto age, x ( in days ) as y = 0.56x + 9.9; RESULTS thus age shouldbe relatedto bill lengthas x = (y-9.9)/0.56 and we have usedthis formulato producea standard For the years 1990-1994, we were able to calculate81 table of age for each half millimetrelength of bill from 10 hatch-dates for the Borders and 99 for the Peak District. mm (age 0) to 21 mm (19.8 days, = 20 days). These rangedfrom year-day146 (24 May) to 196 (13 July), but the mean hatchingdates for the differentyears For the years 1990-1994, we have data for 8-23 broods and sitesclustered in the period6-18 June (year-days from the Borders(mostly Leithen Water and Dewar 159-171). In each year, the mean hatchdate in the Water, southof Edinburgh,collected by T.W.D.) and for Borderswas earlierthan in the Peak District,by 1-5 days, 10-25 broodsfrom the Peak District(mostly the but this differencewas not quite statisticallysignificant LadybowerReservoir system, but also fromthe River (Meddis'H = 2.84, p = 0.088, d.f. = 1). However,the Ashop,collected by D.W.Y.). Forthe Peak District,we differencein hatch-datesbetween years, allowing for any have reworkedthe data for 63 broodsused by Holland& differencebetween sites, was highlysignificant (H =

53 12.81, p = 0.012, d.f. = 4), and moreoverfitted best the Yalden (1991), at year-day167.6 (14-15 June). However, hypothesisthat 1990 was the earliestyear, 1991 was the the hatch-dates for the known nests of Robson eta/. latest,and the otherthree years were equal (Z ---3.51, p < (1982), fromthe 1970s, averagedthree days earlier 0.001). Hatch-dateswere around 10 days earlier in 1990 (Table3), and this resultwas statisticallysignificant than in 1991 (Table 1). (Meddisspecific test, with no a pr/or/expectationof directionof change,Z = 1.985, p = 0.046). Table 1. Hatch-datescalculated for CommonSandpiper chicks in the Peak District and in the Borders, 1990-1994. DISCUSSION PEAK 1990 1991 1992 1993 1994 DISTRICT Despitethe impressionof synchronyin breedingseason Mean hatch 160.5 170.8 168.3 167.0 167.1 from year to year, our resultsindicate the variationin date S.D. 1.7 2.0 2.2 2.2 2.6 average hatch-datebetween years. We have n 10 25 22 25 17 demonstratedthat the growthrates of CommonSandpiper Earliest 152 160 157 150 152 chickswere faster in 1992, a warm, dry June, and slower Latest 173 196 195 188 187 in 1991 whichwas cold and sunless(Yalden & Dougall 1994). Hatchdates were clearlysimilarly affected by BORDERS differentweather conditions,and May was also cool in Mean hatch 159.3 165.3 165.4 163.6 166.3 1991, while it was warm in 1990. date S.D. 3.4 4.9 2.6 1.8 2.7 The indicationthat the breedingseasons were somewhat n 10 8 18 23 22 Earliest 149 146 15O 146 1 50 earlier in the Borders,280 km north of the Peak District, Latest 187 192 185 182 195 was surprisingand counter-intuitive.There is no differencein altitudebetween the studysites (bothrange (Differencebetween years, Meddis H = 12.8, p = 0.012) 700-1000 ft, 210-300 m), so one would expectthe (Differencebetween places, Meddis H = 2.8, p = 0.088) northern site to be later. Such a result could occur if one site (Peak District)were studiedmore persistently than the Althoughthe difference in hatch-dates between the other(Borders); however, the discrepancyis particularly Bordersand Peak Districtstudy sites was not quite the absenceof early (May) hatchdates, at the time when formallysignificant, there were distinctlymore hatch-dates surveyors'enthusiasm is greatest. in May in the Borders,and, conversely,more dates in July in the Peak District;analysed this way, there was a A clue may be given by the somewhatearlier hatch-dates significantdifference between the areas(X 2 = 12.3,p = in the Peak District in the 1970s, which show the same 0.003) (Table 2). mean hatch-date(year-day 164) as the recent Borders sample. Hatch-datesseem to have slippedback in the Table2. Numberof hatch-datesin May, Juneand Julyin the two Peak District around 1980, but to have been stable since study areas, 1990-94 combined. then. This differencemight have been causedby the differencein methodology(known hatch dates, versus May June July Total calculatedhatch dates from chickscaptured for ringing), Peak District 5 83 11 99 but, assumingthat it is genuine,the impactof the late Borders 18 58 5 81 Aprilsnowstorms of 1981 mightbe the cause. We have documentedthe sharpdecline that year of our Peak (•2 = 12.35,p = 0.003) Districtstudy population (from 22 pairsin 1980 to 14 pairs in 1981) (Holland& Yalden 1991b),and we suspectthat Table 3. Variationin mean hatch-date in the Peak District, older,well-established and moreexperienced birds comparingpre-1980 data from Holland et al. (1982)with1980-89 sufferedheavier mortality than youngerbirds, in part data fromHolland & Yalden(1991a) and the recentsample from becausewe expectthem to returnto their territoriesmore 1990-94. promptlyat the startof the season (Holland& Yalden 1995). However,the 1981 snowstormwas a localrather pre-1980 1980-1989 1990-1994 than a nationalphenomenon, and had no effectin the Borders area. Mean 164.7 167.6 167.6 S.D. 1.4 1.1 1.0 Our resultsclearly indicatethe sensitivityof the timingof n 49 63 99 the breedingseason in CommonSandpipers to the Earliest 149 152 150 Latest 198 192 196 weather,just as the growthrate of the chicksis similarly sensitive.The shortbreeding season, particularly its curtailmentin July, suggeststhat it is criticalfor the young (Meddisspecific test, pre-1980 earlier than others, Z = 1.99,p = 0.046) to fledgeas early as possible,either because the food supplyon the breedinggrounds declines through July (Yalden1986) or perhapsbecause the fledglingshave to Withinthe Peak District,the average hatch-datefor the grow further and accumulatefat for their southerly aggregate 1990-94 sample is identicalto that for the migration.Presumably the changeindicates that laying 1977-1989 sampleof chicksmeasured by Holland& dates, and thereforethe start of incubation,are delayed in

54 6 PEAK DISTRICT pre-1980

6 PEAK DISTRICT 1980-89

6t•40 145 !150 •155 160 J165 Ii170 175 BORDERS180 1851990-94 190 195 200 Year - Day Figure1. Frequencydistribution of hatch-datesfor CommonSandpipers Actitis hypoleucos. The hatchdates for 49 knownnests reported by Hollandet al. (1982) ("PeakDistrict pre-1980") are comparedwith dates calculated back from the age of chickscaught for ringingby Holland& Yalden(1991) ("Peak District 1980-89"), caught more recently in the Peak District,and in the Borders.Hatch-dates are givenas year-days(152 = 1 June, or 31 May in a leap year).

"later"years, and this must reducethe opportunitiesfor REFERENCES replacementclutches to be laid if the firstclutch is lost. The tail of later dates (Figure1),those after year-day175 Cuthbertson,E.I., Foggitt,G.T. & Bell,M.A. 1952. A censusof (23 June), is presumedto representthese replacement CommonSandpipers in the Sedbergharea, 1951. Brit.Birds clutches. Presumablythe averagehatch-date represents 45: 171-175. some sortof optimumfor the species,in whichcase the apparentlylater date now (but notformerly) in the Peak Holland,P.K., Robson,J.E. & Yalden,D.W. 1982. The breeding biologyof the Commor,Sandpiper (Actitis hypoleucos) in th• District,and the largertail of July hatch dates in the Peak Peak District. BirdStudy 29:99-110. Districtthan in the Borders,suggest that conditionsin the breedingseason in the Peak Districthave changed. Holland, P.K. & Yalden, D.W. 1991a. Growth of Common Sandpiperchicks. Wader StudyGroup Bull. 62:13-15. This speculationwould be strengthenedby informationon Holland,P.K. & Yalden,D.W. 1991b. Populationdynamics of hatch-dates,growth rates and breedingsuccess of CommonSandpipers (Actitis hypoleucos) breeding along an CommonSandpipers elsewhere in their range. uplandriver system. Bird Study38: 151-159. Cuthbertsoneta/. (1952), workingat Sedbergh,midway betweenour studysites, reporteda mean hatchdate for Holland, P.K. & Yalden, D.W. 1995. Who lives and who dies? The 14 nestsof year-day 166, 13 June (range4-23 June) in impactof severeApril weather on breedingCommon SandpipersActitis hypoleucos. Ring & Migr. 16: (in press) the years 1949-51. This is a smallsample, but coincides well with the other dates. More evidenceon average Meddis,R. 1984. Statisticsusing ranks: a unifiedapproach. dates and their variationfrom year to year would be Blackwell, Oxford. welcome. The BTO's nest recordcards for this species Yalden, D.W. 1986. Diet, food availabilityand habitatselection of have not been analysed,but may not indicatehatch-date breedingCommon Sandpipers Actitis hypoleucos. Ibis 128: with sufficientaccuracy. However,there may be merit in 23-36. examiningthis sourceof data. Yalden,D.W. & Dougall,T.W. 1994. Habitat,weather and the growthrates of CommonSandpiper Actitis hypoleucos chicks. WaderStudy Group Bull. 73: 33-35. ACKNOWLEDGEMENTS

Datafrom the Bordersregion were collectedby Borders RingingGroup and collatedby T.W.D. Recent Peak Districtdata were collectedby D.W.Y., pre-1990 data by both D.W.Y. and P.K.H.. We thank the various land owners,particularly Severn-Trent Water PIc, Dalmenyand RaeshawEstates, for permissionto studyCommon Sandpiperson their land. Dr R.R. Baker ran the statistical teststhrough his versionof Meddis,for whichwe are very g ratefu I.