THE SEX-RATIO AND THE PRODUCTION OF THE MALLARD, ANAS PLATYRHYNCHOS L.

Transcription

1 THE SEX-RATIO AND THE PRODUCTION OF THE MALLARD, ANAS PLATYRHYNCHOS L. by J. A. EYGENRAAM INSTITUUT VOOR TOEGEPAST BIOLOGISCH ONDERZOEK IN DE NATUUR (ITBON) ARNHEM, KEMPERBERGERWEG 11 MEDEDELING NR. 34/1957 Overdruk ut ARDEA", TIJDSCHRIFT DER NED. ORNITHOLOGISCHE UNIE JAARGANG 45, Afl. 3/4. DECEMBER 1957

2 1957] 117 THE SEX-RATIO AND THE PRODUCTION OF THE MALLARD, ANAS PLATYRHYNCHOS L. l ) by J. A. EYGENRAAM (Insttute for Bologcal Feld Research, Arnhem, Holland) I. THE SEX-RATIO a. For a proper management of the Mallard populaton a knowledge of the sex-rato s of prmary mportance. Wthout ths t s mpossble to study the effect of reproducton rates and mortalty on the populaton densty by shootng and decoyng. For ths reason, asssted by voluntary helpers, the author made a study of the sex-rato of the Mallard n the feld and n duck decoys. By means of perodc counts made from the end of July to the begnnng of June n varous habtats t was possble to obtan a pcture of the apparent and actual fluctuatons n the sex-rato of the Mallard. Moreover, ths nvestgaton was carred out at the request of the Internatonal Waterfowl Research Bureau, London and organsed n the Netherlands by the author. For the purpose of makng a study of the sex-rato among duck decoycatches, the help was obtaned of a number of decoymen. Durng the months of August and September they noted the catches daly, specfed accordng to age and sex. After September, when the dstnctve talfeathers are moulted (WITHERBY and co-workers, 1952), t becomes a very laborous task to dstngush between adult and juvenle ducks, and the specfcaton s lmted to sex. By ths method of determnng the sex of young ducks only just ready for flght we ganed nformaton on what s known as the tertary sexrato,.e. the sex-rato on reachng maturty. It should be noted that no study was made of the prmary and secondary sex-ratos,.e. those respectvely before and after the eggs are hatched. The results of the counts and decoy catches wll now be dscussed n successon. b. In order to rule out the possblty of faulty conclusons arsng *) Verschjnt tevens als Mededelng No. 34 van het Insttuut voor Toegepast Bologsch Onderzoek n de Natuur.

3 118 [Ardea 3/4 from a sexual dfference n habtat preference, the counts were made n dfferent localtes. The observaton areas are shown n Fg. 1. Fg. 1 Observatons were usually made twce a month, over the perod All round-fgure estmates were excluded from the materal collected, only actual counts beng employed. A total of 865 counts were worked out, dvded nto two groups, vz.

4 1957] 119 cty parks and canals (441 observatons), and more natural envronments such as lakes, pools and rvers (424 observatons). The dvson was made throughout because n the towns the Mallard s to some extent farly tame, occasonally domestcated, and durng the wnter months ts numbers are scarcely or not at all renforced by mgratory Mallard. In natural surroundngs t s manly the true wld Mallard that s encountered. Should there be any queston of a separate mgraton of the sexes, ths could be more clearly demonstrated n the latter localtes than n the former, and mght affect the result when the data are combned. In both cases, however (Tables 1 and 2 respectvely) the seasonal fluctuatons show the same trend. Month Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mrch Apr. May June Total TABLE 1 Sex Rato n Mallard 1948/1954 Numberof observatons In Cty parks and canals Drakes Ducks <J<? : 100?? From August to January there s frst a large, and afterwards a gradually decreasng majorty of drakes. From March onwards there s agan a rapd ncrease n the excess of drakes over ducks, reachng ts hghest pont n June. Snce the fluctuaton trends n Tables 1 and 2 are n agreement there s no harm n combnng data from all areas and calculatng the averages from these fgures (Table 3, Fg. 2). Let us now consder these averages more fully.

5 120 [Ardea 3/4 TABLE 2 Sex rato n Mallard 1948/1954 On pools, lakes and rvers Month Number of observatons Drakes Ducks 33 :100?$ Aug. Sept. Oct. Nov. Dec. Jan. Feb. March Aprl May June Total TABLE 3 Sex rato n Mallard 1948/1954 Total average Month Number of observatons Drakes Ducks 33 : 100?? Aug. Sept. Oct. Nov. Dec. Jan. Feb. March Aprl May Tune Total

6 1957] 121 It s notceable that both n sprng and autumn the drakes are greatly superor n numbers. The excess number of drakes n the autumn may be accounted for by the fact that the wng moult, whch makes the adult Mallard for some tme ncapable of flght, begns earler n the male and s thus also completed earler than n the female. Durng ths "flghtless" perod most Mallards lve at least a month n localtes affordng a good deal of cover whch they generally only leave n the evenng n order to go foragng on foot or swmmng. For ths reason they escape observaton. By August most drakes have / _ apparent fluctuaton actual fluctuaton / t - ^ \ \ L ] VI) IX XII IV VI Fg. 2 completed moultng ther wng-feathers, whereas the ducks have only just begun (LEBRET 1949). Hence the ducks beng less n evdence, the drakes are only apparently very much n the majorty. The ducks dsappear agan n the sprng, but ths tme n order to lay and ncubate the eggs. Afterwards they rear ther broods and lve separated from the drakes. It s therefore mpossble to arrve at a correct estmate of the sexrato of the Mallard from the begnnng of the breedng season untl the end of the moultng season. The latest date on whch an exceptonal female Mallard was encounter-

7 122 [Ardea 3/4 ed wthout ts wng-feathers s the mddle of October (communcated by MAEBE and VAN DER VLOET). The dffculty n calculatng the real sex-rato n the feld, caused by the factors mentoned, ceases after October. The only factor whch could then have a dsturbng effect would be a sexually dfferentated mgraton. Should ths phenomenon be present to a hgh degree among Mallard from northern Europe t s concevable that one sex may cross or wnter n our lattude n greater numbers than the other. Ths would also make t dffcult to dscover the real sex-rato durng the wnter months, the Dutch Mallard beng predomnantly of a resdent nature (LEBRET 1956). The queston as to whether there s a sexually-dfferentated mgraton of Mallard may best be studed from the daly catches n duck decoys. Durng the perod from rushes of the northern Mallard were repeatedly caused by the dvson of temperature n western Europe, vz. from md-december 1950 to the begnnng of January 1951; n november 1953 and January 1954; n November 1954 and January At these tmes there were sudden ncreases n the catches, followed by a gradual fallng-off. Durng these perods of rush-mgraton, whch comprse a total of 85 days, 9725 Mallard were caught n certan decoys n a rato of 95 to 96 drakes: 100 ducks. Outsde these mgraton perods Mallard were caught n the same decoys n the months of November, December and January n a rato of 119 drakes: 100 ducks. It s clear that durng the rushes the females are n a slght majorty, and at other tmes the drakes. The females are apparently more senstve to frost whch causes them to start mgratng somewhat earler, but t s obvous that the drakes follow soon after. Ths can be shown by the average sex-rato of all of the catches durng the months of November, December and January over the years as compared wth the average sex-rato n the feld. In the decoy catches ths average s 108 to 109 drakes: 100 ducks. The average of the feld counts over the same perod s 109 to 110 drakes : 100 ducks. Ths strkng agreement shows that the duck decoys do not trap selectvely durng ths perod, and also that the dfferentaton n mgraton only results n a very temporary dsturbance of the sex-rato. Hence t has no vsble effect on the perodc counts. From the decoy catches we may assume that the sex-rato among young brds August-September s 113:100 (Table 4). It s worth notng that ths fgure s relable because juvenle ducks do not moult ther wng feathers

8 1957] 123 TABLE 4 Sex-Rato n Mallards Caught n Duck Decoys <J<? ad.?9 ad. (?<?juv.?juv. Total Aug. Sept. Total Sex rato Aug. Sept. Total Sex rato Aug. Sept. Total Sex rato Aug. Sept. Total Sex rato Aug. Sept. Total Sex rato Total general Averagesexrato :: :: :: : :: :: : : :: :: n the frst year and duck decoys do not catch selectvely drakes or ducks. Among the total Mallard populaton (juvenles and adult brds) the rato s 114 n November (Table 3). Later t gradually falls off to about 106:100 n January. After ths month nvestgaton of the decoy catches cannot be contnued because the close season begns on the frst day of February. The ndcatons are that n ths case we are concerned wth a real fluctuaton, snce nether the moultng of the wng-feathers nor the dfferentated mgraton can affect the result. Nor are the decoy catches a factor as they do not work selectvely after October. They are, however, selectve before November for the adult brds (Tables 4 and 5) as the females moult durng the huntng season, unlke the drakes, and cannot be caught.

9 124 [Ardea 3/4 TABLE 5 Sex Rato n Mallards Caught n Duck Decoys Males :100 females Oct. Nov. Dec. Jan. 1950/ / / / / Average Numbers of ducksnv There s thus a shorter perod n whch they are lable to be captured and as a result they pay a lesser toll. But ths s reflected n the catches up to and ncludng October, and not n those from November to January. As shootng of ducks durng the wngmoult s forbdden n our country t s lkely that durng the perod of August to November sportsmen wll shoot more drakes than females. It s possble that the gradual fallng - off from 114 n November to 106 n January s also caused by selectve shootng durng these last months of the huntng season (BOYD 1954). Ths s a matter for further nvestgaton. Summary and conclusons 1. An nvestgaton was made nto the sex-rato of the Mallard n the Netherlands by means of smultaneous feld counts and duck decoy catches. 2. Snce the perod of the wng-moult does not concde n the drake and the duck t s practcally mpossble to carry out an nvestgaton nto the sex-rato before November. 3. From July to November huntng (ncludng decoyng), whch s the chef mortalty factor, has a more marked effect on the drakes than on the ducks, the latter beng less vulnerable for a tme. 4. The sex-rato n November s 114 : 100. In summer t wll therefore probably be hgher among adult brds as by selectve shootng and decoyng n the autumn t has been forced down to ths fgure.

10 1957] Durng the wnter there s a further, albet small declne whch s possble caused by selectve shootng alone, decoy captures beng not selectve n wnter. 6. Immedately pror to the breedng season the sex-rato s approxmately 106 : It follows that the sex-rato s not statc (LEBRET 1949) but dynamc and vares n the course of the year between 106 and an unknown fgure exceedng 114. Samenvattng Om de populatedchthed van een vogelsoort te kunnen beheersen s o.m. kenns omtrent de geslachtsverhoudng noodzakeljk. Voor de Wlde Eend s deze bestudeerd aan de hand van een groot aantal waarnemngen n verschllende botopen. Ook kon de geslachtsverhoudng n de vangsten van eendenkooen worden vergeleken met de n het veld. De perode, waarn deze stude kan worden utgevoerd, wordt door verschllende factoren beperkt. Doordat de slagpenru bj woerden en eenden net geljk plaats vndt, vallen de peroden, waarn de woerden en de eenden zch weng vertonen, net samen. Tjdens het broedsezoen, dat n maart begnt, vertonen de wjfjes zch mnder dan de woerden. In verband met deze bede verschjnselen kan de werkeljke geslachtsverhoudng bj de Wlde Eend n het veld slechts worden bepaald van november tot en met februar. Het resultaat zou ook dan nog net betrouwbaar kunnen zjn, nden de Wlde Eend een geslachteljk gedfferenteerde trek zou vertonen. Dt verschjnsel s er nderdaad, zoals ut de koovangsten kan worden dudeljk gemaakt. Bj rush"-trek komen de wjfjes voorop, maar doordat de woerden weldra volgen, s de verstorng van de geslachtsverhoudng van zo korte duur, dat de gevolgen n de veldtellngen net werden waargenomen. Bj de Wlde Eend wordt steeds een klene meerderhed van woerden vastgesteld. In november bedraagt deze 114 ó*ó* tegen 100 $?, n februar nog 106 : 100. Deze vermnderng zou kunnen worden veroorzaakt door selectef afschot, wat nog nader zal worden onderzocht. Het s net waarschjnljk, dat de eendenkooen selectef vangen, omdat de verhoudng woerden : eenden n de koovangsten gedurende de genoemde perode geljk was met de n het veld. Tussen februar en november s de geslachtsverhoudng slechts bj benaderng te schatten. Ut de omstandghed, dat n november, d.. na een perode van dre maanden waarn de druk van jacht en koovangst op de woerden groter s dan op de eenden, de verhoudng woer- Ardea, XLV 9

11 126 [Ardea 3/4 den : eenden nog 114 s, moet worden afgeled, dat deze vóór de openng van de jacht nog hoger s geweest en door het eenzjdg scheten en vangen van woerden tot deze waarden s teruggebracht. In februar zjn er slechts 106 woerden per 100 wjfjes en dt suggereert, dat tengevolge van een grotere sterfte onder de wjfjes gedurende het broedsezoen de sekseverhoudng oploopt tot boven 114. In verband met het fet, dat de woerden de gehele zomer gescheden leven van het overgrote deel der wjfjes, n.1. van de wjfjes de jongen hebben, zal de dynamek n de geslachtsverhoudng zeer moeljk n cjfers kunnen worden bestudeerd. Het s waarschjnljk, dat de geslachtsverhoudng zch n de loop van de tjd beweegt van 106 : 100 n februar va een waarde hoger dan 114 naar 114 : 100 n november. Het onderzoek was mogeljk door de zeer gewaardeerde medewerkng van een aantal waarnemers, welker namen her met erkenteljkhed worden genoemd. Mej. G. v. D. BAAN, mej. A. KOETEN en de heren J. ALBERS, P. ANNEMA, O. BAJEMA, A. v. D. BERG, R. v. D. BERG, A. BER- KOUWER, A. M. BLOKLAND, SJ. BRAAKSMA, P. W. BRANDER, PH. G. GOOSSEN, G. GOOYER, C. GROOT, C. HOOGERHEIDE, A. DE KLEUVER, J. A. F. KORIDON, T. LEBRET, J. MAEBE, J. G. M. MARQUENIE, H. T. v. D. MEULEN, H. P. MISSET, J. G. J. M. MOOYMAN, M. F. MÖRZER BRUYNS, G. NOOTER, K. POSTMA, M. W. STRUYCKEN, G. SUKKEL, C. SWENNEN, A. TIMMERMAN, A. A. TJITTES, J. VIERGEVER, H. V. D. VLOET, J. W. DE VRIES, J. WALTERS en een aantal kooleden. Lterature cted Boyd, H., 1954: Some Results of recent Brtsh Mallard Rngng. The sxth Annual Report of the Wldfowl Trust London. Frelng, H., 1934: Statstsche Untersuchungen über das Geschlechtsverhältns der Enten zur Zugzet. Der Vogelzug 5, pp Lebret, T., 1949 : Eenden tellen. De Nederlandse Jager 54, pp. 4-5 ; 26, pp Lebret, T., 1956: De Wlde Eend, Anas platyrhynchos (L)., als Nederlandse Standvogel. Ardea 44, pp Mc Ilhenny, A., 1940: Sex rato n Wld Brds. The Auk 57, pp Wtherby, H. F. as., 1952: The Handbook of Brtsh Brds. Vol. Ill, p II. THE PRODUCTION The commonest method of determnng the reproductve success of a duck speces s actually to keep a number of ndvdual brds undej observaton for an entre breedng season n a compact and well-defned

12 1957] 127 area (BENNETT 1938, EARL 1950, SMITH 1953). On the bass of the results obtaned n the observatonal area the ncrease n populaton s estmated n a much larger area, and possbly over an entre state. The relablty man observaton areas of mallard broods Fg. 3 of ths method wll wholly depend on whether the observatonal area s a representatve one. The author's man reason for adoptng a dfferent method of studyng the producton of the Mallard was that no observatonal area can be found n Holland that may be consdered representatve of the country as a whole.

13 128 [Ardea 3/4 Ths method conssted n carryng out observatons wth the ad of a number of voluntary helpers nto Mallard broods at many places from Aprl to August. The observatonal areas were spread over all parts of the country sutable for Mallards (Fg. 3). The number of samples and the places where they were taken are a suffcent guarantee of the materal beng representatve. In ths way 268 dfferent Mallard broods were counted n 1950; 662 n 1951; 620 n 1952; 990 n 1953 and 1177 n a total of 3717 dfferent Mallard broods. Each year some tens of broods were seen by the same observer more than once wth ntervals of at least a week, thus ncreasng the number of data to There s no ndcaton that reports made by dfferent assstants relate to the same broods. The date, locaton, number of ducklngs and ther age were noted n the case of every Mallard brood observed. For estmatng the age crtera were gven whch were even more extensve than those employed by SOUTHWICK (1953), so that n the case of very small ducklngs the age could be estmated n days, and n the case of all other ducklngs n weeks. Wth the use of the nformaton regardng date of observaton and age t was possble to calculate the approxmate date of hatchng n the case of each brood. The average sze of brood per agegroup could be calculated from the nformaton on the number of chcks per brood and the estmated age. The producton was calculated from the entre materal n combnaton wth the sex-rato and the percentage of unproductve ducks. Dstrubuton of the hatchng dates The Mallard s able to start layng from February to September. The ncubaton perod s 28 days and t takes 12 days to produce a full clutch of 12 eggs, so that the Mallard chcks may hatch about 40 days after the frst egg has been lad. In some years the frst ducklngs appear before 1st Aprl. Durng ths month the numbers ncrease steadly untl n the frst decade of May a clmax s reached whch occurs year after year, but s more marked n some years (1951 and 1954) than others (Table 6). TABLE 6 Dstrbuton of hatchng dates on a percentage bass *) Year Apr. I Apr. II Apr. Ill May I May II May III June I June II June III x ) The very early and the very late broods have been omtted

14 1957] 129 Second broods are very exceptonal, but havng regard to the many broods that are stll found after 31st May there must be many cases of repeat clutches. The latter generally number from 7 to 11 eggs. Hence the June broods are smaller than the Aprl and May ones (Table 7). Mallard ducklngs may also be born as late as July, August TABLE 7 Relaton between brood-sze and hatchng perod Average brood sze of ducklngs one day old Dtto on the age of 5 weeks and older Hatched n: Aprl May June Aprl May June Average or even September, but the very early and very late broods are of lttle mportance for the mantenance of the speces. In 1950 the number of broods hatched before 1st Aprl and after 30th June was 4% of the total, n %, n %, n %. Consequently at least 94% of the Mallard ducklngs are hatched durng Aprl, May and June (Table 6) The average brood-sze Durng the frst few days of ts lfe a brood averages nearly 11 chckens. The Aprl and May broods may have a rather hgher average and the June [and July] broods a rather lower one (Table 7). The mortalty durng the frst week as a result of nsuffcent vtablty, unfavourable weather condtons and prédaton causes a rapd declne n ths number to the averages of between 7.44 and 9.49 as shown n table 8. Year TABLE 8 Average brood-sze per age group 0-1 wk. 1-2 wk. 2-3 wk. 3-4 wk. 4-5 wk. 5-6 wk. 6-7 wk. 7-8 wk

15 130 From the frst to the second week of lfe there s a further sgnfcant drop, but subsequent losses are apparently slght (Table 9). Actually TABLE 9 Numbers of broods from whch the data were calculated Year 1 wk. 2 wk. 3 wk. 4 wk. 5 wk. 6 wk. 7 wk. 8 wk. Total Ill Total they are more consderable than mght be nferred from the consecutve columns. Ths s owng to the fact that a number of broods persh altogether and are not allowed for n the averages; they wll nevertheless have to be ncluded n one way or another n the fnal result. Ths s done by estmatng the percentage of unsuccessful females (p ). There are ndcatons that very small broods of one to two ducklngs occur to a comparatvely less extent n the older age groups than n the younger ones. It seems, therefore, that broods reduced to less than three have less chance of survval. The probablty of ths can be demonstrated by analysng the numercal frequency of the broods n the varous age groups. It was found that n the age group 1 (0-1-week-old ducklngs) the number of ducklngs vared from one to 19 per brood and that the most frequently occurrng number was ten (Fg. 4). In the age group 2 (1-2- week-old ducklngs) the brood-sze vared from one to 15, seven beng the number occurrng most frequently (Fg. 5). The varaton and frequency trend n the other groups can be seen from Fgs To facltate a study of the dstrbuton of brood-szes varyng from one to 12 n all eght age groups ths s summarsed n Fgs The large broods,.e. those comprzng 9, 10, 11 and 12 ducklngs, are, on a percentage bass, consderably more numerous n the younger age groups than n the older ones: the graph shows a steeply downward trend (Fg. 12). The mddle groups (broods of 5, 6, 7 or 8 ducklngs) show a proportonate ncrease n numbers, whch s very understandable when one consders ther contnual renforcement from orgnally larger broods reduced by mortalty to broods of 5-8 (Fg. 13). The small broods of 3-4 also

16 -15% -10% -5% 0-1 weeks Fg. 4-15% -10% -5% 1-2 weeks 5 IO Fg. 5 15

17 -15% -107» weeks 5 IO Fg weeks 5 IO Fs. 7 I5

18 % 5% 4-5weeks I 5 IO I5 Fg. 8-15% -10% 5% 5-6 weeks I 5 IO I5 Fg. 9

19 -15% -10% 5% 6-7weeks I 5 IO I5 Fg % -10% 5% 7-8 weeks 5 IO I5 Fg. 11

20 135 15% lov IO% 5% I I I I I I L I weeks Fg. 12 show an upward trend, and ths s to be explaned n the same way as n the case of the medum-sze broods. But n the case of the very small broods of 1 and 2 we agan notce a downward trend, whch means that on a percentage bass the latter occur less frequently n the older age groups than n the younger ones (Fg. 14). Ths reducton mght be due to the fact that a normal loss s only made good from above to a very slght extent. Ths would mean that the parental nstnct of a duck that has managed to retan three or four young works so effectvely that

21 136 [Ardea 3/4 15% IO% 5% weeks Fg. 13 further losses are an nfrequent occurrence. Ths theory would be supported by Fgs. 8, 10 and 11, n whch the transton from three to two s very steep. In Fgs. 4, 5, 6, 7 and 9 the transton s so gradual, however, that t s unlkely that lttle mortalty occurs n broods of three and four ducklngs. It s much easer to assume that n these broods also the

22 1957] % / * /' \ / \ V \ lor. 5% -I l_ we<?ks Fg. 14 mortalty s such that the numbers of broods wth one and two chcks are regularly renforced. Should they nevertheless declne n number, the losses are bound to exceed the gans. In ths case the chances of very small broods survvng must be small. The dsappearance mght be caused by the ducklngs leavng the mother to jon other broods, or by the mother abandonng her chcks. The frst assumpton s mprobable, frstly because out of the 3717 broods

23 138 [Ardea 3/4 on whch we have nformaton only n 0.1% of the cases t was reported that the brood conssted of chcks of varyng ages. Feld observatons also show that there are practcally no cases of adopton after the frst day of lfe. Both the mother duck and the young are hostle to ducklngs from another brood. What presumably happens n these cases s that the ducks forsake the chcks as soon as only one or two are left and the latter soon persh. If the season s stll not too far advanced and a number of sexually potent drakes reman n the vcnty, a duck of ths knd mght start producng a repeat clutch, and ths would serve the speces better than tendng a few chcks. However, the accuracy of ths explanaton needs to be tested by experment., Whatever the cause may be, the numbers of broods consstng of one and two ducklngs are relatvely smaller n the older age groups than n the younger ones and therefore the averages calculated from the observed broods of older age are too hgh and have to be corrected n order to determne the producton. In other words we have to know the percentage of unproductve females. Reproductve success It can be seen from the last column n Table 5 that the number of ducklngs per successful duck reachng maturty n the fve years of the nvestgaton s practcally the same. The dfference between the extremes (6.02 and 6.91) s not sgnfcant. Where the number of fully-grown young per successful duck s constant, good and bad breedng years must be determned by means of three factors : 1. The number of nestng brds; 2. the annual dfferences n mortalty shortly after attanng to maturty. 3. the percentage of unproductve females. These factors wll now be dscussed n turn. 1. No exact data are n hand concernng annual fluctuatons n the number of nestng Mallards n the Netherlands. On the bass of wdespread counts, the nestng brd populaton n 1956 was estmated at (OVER en MÖRZER BRUYNS 1956)*), less than half ths number beng females. It s lkely that ths number fluctuates to a certan extent. 2. The perod between beng old enough to fly and the openng of the shootng season on 24th July s short. There were no sgns of any *) In the duck decoys another Mallards were present at that tme (Author).

24 1957] 139 catastrophc mortalty n July and August of the years covered by the nvestgaton, so that ths possblty may be dsregarded. 3. The percentage of unproductve females may vary, although t s true that no such dfference can be shown n the two years n whch we have suffcent relevant nformaton. To obtan some dea of ths percentage t s necessary to determne the sex-rato at a perod n whch the successful females stll accompany ther chcks and the ducks that have lost ther eggs or young do not subsequently begn on a repeat clutch,.e. the frst decade of June. Strctly speakng ths s not qute correct for the followng reasons: a. Mallard chcks are seen as early as the thrd decade of March. They are ready to fly after 56 days, and wthn a further fortnght the parent duck s able to leave them. Durng the frst decade of June the frst successful females of the season are able to jon the drakes lvng a socal lfe. b. A female that begns to lay on 15th June may stll have chcks 25th July, and snce ducks wth newly hatched ducklngs are stll seen n August, or even later, the ducks counted between 1st and 10th June may nclude ndvdual brds that are stll begnnng a repeat clutch. There s thus an overlap and the resultant error must be accepted. In practce, however, the drawbacks referred to here are much less serous, frstly because only a small percentage of ducks have chcks even before 1st Aprl and the brood s moreover very lable to persh, and secondly because the June counts are n respect of groups of Mallards that are sexually ndfferent. It s not lkely that females stll ready to par wll seek the socety of such drakes. For these reasons t may be assumed that the error s only theoretcally possble. For beng able to make an estmate as accurate as possble of the percentage of unproductve females n the frst decade of June t s nessesary to know the sex rato at that tme. We only know the sex rato n February to be 106 males : 100 females and ths n November to be 114 : 100. Presumably n June the drakes outnumber the females even more than n November (see page 124) and therefore a sex rato of 115 to 125 seems a justfed estmate. The counts dscussed here were made wth some success n the frst half of June 1953 en In Mallards were counted n 11 locatons n a rato of 577 0*0* to 106 ÇÇ. In 1954 n the same 11 locatons 1085 Mallards were counted n a rato of 929 <$3 to 156 ÇÇ. Ths tme the count was made on a larger scale vz. at 41 locatons, gvng a total of 4824 Mallards n a rato of 4059 ctc? to 765??

25 140 [Ardea 3/4 Acceptng a sex rato of we mght expect for the number of 577 SS n 1953 a number of ÇÇ. There were only 106. Hence 20-23% was unproductve. In 1954 for 929 drakes 156 ÇÇ, resp. for 4059 &J 765 ÇÇ were counted. For these numbers $$ resp ÇÇ mght be expected. There were only 156 resp So 19-21% or 22-24% was unproductve. The dfferences are not sgnfcant. The 1953 producton was thus 6.61 (Table 5 last column) 20 to 23% = 5.09 to The 1954 producton was to 24% = 4.88 to Thus n both years about 5 ducklngs per duck reached maturty SA 1955 catches tn ~^-v / 3 monfhs ',okf Ç 50 5! ,4 55 Fg. 15 For a populaton of nestng Mallards of whch were females the reproducton for 1953 and 1954 would have been Fnally, we wll also deduce whether these years were good, moderate or poor by analysng the decoy-catches n the months of August, September and October. These fgures have been avalable snce 1949,.e. the year after the waterfowl research was begun. Fg. 15 shows these catches n the form of averages per decoy. On studyng the graph we notce at once that the August catches beng substantally the same year after year, are unable to provde any nformaton on the success n breedng. There are no apprecable dfferences untl September and October. In vew of the fact that n some years mgratory Mallards may arrve n numbers as early as October t s better to leave

26 1957] 141 ths month out of account. We are thus left wth September only. In both years the catches n ths month were below the average, but otherwse show the same amount of agreement as the calculated percentages of unproductve ducks. Ths suggests that n years of hgher September catches (1950 and 1955) ether the percentage of unproductve brds was less than 20, or the number of breedng brds was hgher than normal. Summary and conclusons 1. The Mallard's producton was nvestgated by counts of Mallard broods, the locaton, date, number and age of the ducklngs beng noted. 2. The hatchng dates were nferred from the data, and a maxmum for the hatchng of young was found n the frst decade of May. 3. The average number of ducklngs reared by a successful female s farly constant and amounts to sx. 4. The percentage of ducks that are unsuccessful n rearng young s probably varable, and was 23 and 24 n 1953 and 1954 respectvely. 5. Ths percentage must be subtracted from the fgure of sx n order to arrve at the reproducton fgure. 6. Judgng from the average decoy catches n September these were by no means favourable years. 7. The nestng populaton may be estmated at Mallards, ncludng females. These rear an average of fve ducklngs, so that the annual producton should be put at Samenvattng Het broedsucces van de Wlde Eend s bestudeerd door het enge jaren volgehouden tellen van eendentomen. Van een groot aantal waarnemers werd hern medewerkng ondervonden. Bj edere waarnemng, werden behalve de plaats, de datum, de leeftjd en het aantal kukens vastgelegd. Ut deze gegevens kon de spredng van de geboortedata, de gemddelde toomgrootte en n combnate met de geslachtsverhoudng het broedsucces worden berekend. Ieder jaar werd mnstens 94% van de eendenkukens geboren tussen 1 aprl en 30 jun, met een mn of meer dudeljke clmax n de eerste 10 dagen van me (Tabel 6). Bj de geboorte tellen aprl- en me-tomen gemddeld ongeveer 11 kukens, de jun-tomen ets mnder, doordat deze meest van vervolglegsels afkomstg zjn (Tabel 7). Ardea, XLV IO

27 142 [Ardea 3/4 De sterfte s n de eerste week het grootst. Ook n de tweede en de derde week zjn de verlezen nog aanzenljk. Afgaande op de gemddelden zou men de ndruk kunnen verkrjgen, dat er daarna weng eendenkukens meer sterven, doordat de gemddelden nagenoeg geljk bljven. In werkeljkhed zjn er wel degeljk verlezen, maar doordat het percentage zeer klene tomen van 1 en 2 stuks onder de jongere leeftjdsgroepen hoger lgt dan bj de oudere, worden de laatste gemddelden mnder gedrukt. De schrjver heeft de ndruk gekregen, dat de overlevngskans van tomen de mnder dan 3 stuks tellen, gernger s dan van numerek sterkere tomen. De zeer klene tomen gaan mogeljk te gronde doordat de moeder hen verlaat. Verlaten jongen worden door andere tomen zeer zelden geadopteerd. De eenden, de er net n slagen een broedsel groot te brengen, sluten zch aan bj de n troepjes levende woerden. In de eerste 10 dagen van jun werden deze geteld. Door de geslachtsverhoudng n deze perode te schatten op woerden : 100 eenden kon het percentage netproducteve eenden worden bepaald. In 1953 bleek dt 20-23% te zjn, n %. In bede jaren berekten per eend ongeveer 5 kukens de volwassen leeftjd. Doordat het gemddeld aantal volwassen geworden kukens per eend constant s (tabel 8, laatste kolom), hangt het broedsucces af van de omvang van de broedvogelpopulate en van het percentage net-producteve eenden. Het s waarschjnljk, dat deze bede grootheden van jaar tot jaar wsselen, hoewel het laatste n de onderhavge stude net het geval was. Er mag rekenng mee worden gehouden, dat n ons land per jaar tenmnste à jonge Wlde Eenden het vlegvlugge stadum bereken. Met erkenteljkhed voor hun medewerkng worden her de namen van de waarnemers vermeld: Mej. G. v. D. BAAN, mej. A. KOETEN. J. ALBERS, P. ANNEMA, F. J. APPELMAN, O. BAJEMA, D. BAKKER, A. V. D. BERG, R. v. D. BERG, E. Bos, G. BOSCH, TH. v. D. BOSCH, K. A. BOSMA, K. BOSMA, P. W. BRANDER, V. BRANDWIJK, G. A. BROUWER, J. C. BRUSSELAARS, J. DIENDER, A. H. v. DIJK, H. EIKELBOOM, B. FENTHUR, M. GEENE, J. V. GEMERT SMITS Jr., W. A. GUNN, N. L. DEN HAAK, G. A. V. HERK, H. W. V. HOYTEMA, M. HOOGENDAM, A. J. HOOGLAND, B. J. HOOGERS, TJ. H. KINGMA, E. v. KOERSVELD, W. A. F. L. KOOLWIJK, J. A. F. KORIDON 3

28 1957] 143 J. TEN KLOOSTER, J. L. TEN KLOOSTER, V. LANGENHOFF, R. LATHOUWER, T. LEBRET, M. v. LENT, H. LINDEBOOM, H. T. V. D. MEULEN, G. MOEDERS- HEIM, A. K. J. MULDER, J. NOUWEN, P. NIJHOFF, L. J. ODINGA, J. H. PEN, K. POSTMA, A. REUSER, H. ROOK JR., A. SCHOORL, A. A. SCHOUSTRA, D. SCHURER, C. v. D. STARRE, C. SWENNEN, J. R. A. TANGE, A. TAMIS, J. J. C. TANIS, B. TUKKER, B. V. D. VEEN, G. VENEMA. R. VISSER, E. E. V. D. VOO, H. J. H. VOSKUIL, D. M. DE VRIES, K. WALDECK, B. J. J. R. WALRECHT, J. WALTERS, B. WESTERHUIS. Staatsbosbeheer, Afdelng Natuurbeschermng en Landschap: M. F. MÖRZER BRUYNS; SJ. BRAAKSMA, C. HOOGERVORST, C. G. v. LEEUWEN, H. P. MISSET, H. J. W. SCHIMMEL, J. TH. DE SMIDT. Lterature cted Bennett, L. J., 1938: The blue-wnged teal. Collegate Press, Ames, Iowa. 144 pp. Earl, J. P., 1950: Producton of Mallards on rrgated land n the Sacramento Valley, Calforna. Jour. Wldl. Mgt. 14, pp Henroth, O. und M. Henroth, 1928: De Vogel Mtteleuropas. Ill Band. Hugo Bermühler Verlag, Berln, p Hochbaum, H. A., 1944: The canvasback on a prare marsh. Amercan Wldl. Inst., Washngton, D. C. 201 pp. Over, H. J. and M. F. Mörzer Bruyns, 1956: Waterwldconcentrates n Nederland n februar De Levende Natuur 59, 7, pp Smth, R. H., 1953: A study of waterfowl producton on artfcal reservors n Eastern Montana. Jour. Wldl. Mgt. 17, pp Southwck, C, 1953: A system of age classfcaton for feld studes of waterfowl broods. Jour. Wldl. Mgt. 17, pp. 1-8.

29