Growth parameters in juvenile Cormorants Phalacrocorax carbo sinensis at Eskilstorps holmar, Skåne 2003 - 05

by Christer Persson and Per Stenberg

Christer Persson, Ljungsätersvägen 43, S-236 41 Höllviken, Sweden. E-mail cp.hollviken'at'swipnet.se
Per Stenberg, Vikensvägen 31 B, S-236 41 Höllviken, Sweden. E-mail per_stenberg'at'tele2.se


Straight text version for printing.

Abstract. Growth parameters of young Cormorants were studied at islets Eskilstorps holmar in the Öresund area, S. Sweden, in the years 2003 - 2005. The study aimed at a better understanding of the dynamic development (with sudden "leaps" in behaviour and performance) of Cormorant populations in a diachronic perspective, analysing first how events in the immediate environment of a colony affect breeding outcome and colony development over a sequence of years. In our case such events of importance to Cormorants have been the scarcity of some fish prey, indicated by poor breeding results of terns in South Scania, and the ongoing culling of Cormorant colonies in neighbouring Denmark, including the termination of attempts to establish new ground-breeding colonies at e.g. Vresen (Great Belt, Denmark) and Saltholm (Öresund, Denmark) since at least 1994. The major result of persecution has been dispersal, bringing together birds with very diverse experiences; it is our belief that culling and food shortage in recent years have interacted to expand the spectrum of dynamic response in both Cormorants and their predators.
The most crucial observation from the biometry collected concerns the isometric relation weight : body size (the latter measured by wing-length); it prevails at least in the wing-length interval 50 - 200 mm, with only a slight falling-off for wing-lengths > 200 mm. A separate analysis reveals that there is a parallel decrease in both weight and wing-length growth rates with increasing size, this state of things being particularly well suited for modelling with the Bertalanffy growth equation: our interpretation of the result is that a year-specific asymptotic weight is aimed at and conserved throughout the nestling period. With the net result of the subtraction [metabolism/anabolism - catabolism] being highly variable, the regulation of juvenile growth must be dynamical to keep growth on the track, the resulting path best described by some sort of a linear attractor (cf. the narrow ellipses of the scattergrams, Figs. 1 and 2, and Figs. 5 and 6). The true distribution of the asymptotic weight is bimodal, and its mean value is likely to be influenced by variations in the sex ratio, if such occur.
Birds controlled at two separate occasions increased their weight at a significantly lower rate in 2004 than in 2005; at the same time the smaller difference in wing growth rate was barely significant at the 5 % level. A period of windy weather may have had adverse effects in 2004, but the quality of food most likely comes into play as well; we believe that the birds fed to a greater extent on freshwater prey with relatively low fat content in 2005.
The breeding success, measured as the number of birds per brood at ringing, varied between 1.94 and 2.18 young/successful brood, only slightly less than North European values from mature colonies in recent years. A novel, surprising feature was heavy gull predation in 2005; gulls pulling out young from the nests - predation as far as we know coinciding with human disturbance in broad daylight - and devouring their stomach contents. At least sixty such cases were recorded. The colony was abandoned in 2006, some sort of disturbance must have been involved.

Contents

1. Introduction
2. Material and methods
2.1. The overall material
3. Results
3.1. The overall material
3.2. Brood size, development within broods. effect of ranking
3.3. Individual development, daily growth rate, asymptote
4. Discussion
Summary

1. Introduction

All factors promoting or constraining population growth in the Cormorant Phalacrocorax carbo sinensis have not been studied in detail, and the synergy effects of interaction between them are little known. On the whole they may be concluded from general considerations, pertaining to a top (or next-to-top) predator of a food-chain. Apart from the self-evident negative effects of various kinds of predation (by birds: e.g. Quintana & Yorio 1998, Volponi 1999, Frere & Gandini 2001, Bregnballe & Eskildsen 2002, by humans: e.g. Bregnballe & Eskildsen 2002, Engström 2001, REDCAFE 2002), disease (Glaser et al. 1999) and pollution (Boudewijn & Dirksen 1995), interaction with other species (Gorski et al. 1990, Gorski & Pajkert 1996), the timing of egg-laying (Debout et al. 1995, Pajkert & Gorski 1996, Stempniewicz et al. 2000) and the degree of synchronization (e.g. Van Rijn 1998) within a colony are of importance, as is the availability of particular prey (Debout et al. 1995, Dirksen et al. 1995, Warke & Day 1995, Suryan et al. 2002, Lorentzen et al. 2004) - the latter factor also involving the turbidity of water (Van Rijn & Platteuw 1996, Van Rijn & Van Eerden 2003) and diving depths (Grémillet & Wilson 1999).

The special feature issue of Ardea (83:1) presents much valuable material, but is hampered by recurrent Darwinist ornaments, in this respect it has been our negative inspiration. We doubt e.g. the lifetime reproductive output (cf. Platteuw et al. 1995) as a useful guiding star for any investigation of reproductive performance in such a complex system. The whole approach automatically ends up stating that the observed reproduction rate is optimal, resulting in the traditional and much celebrated Darwinist tautology. The time-lag of regulation in a species like the Cormorant allows both more-than-optimum and less-than-optimum reproduction over extended periods; one and the same dynamic system tolerating a wide spectrum of starting parameters. One generation may be gaining the experience, the next one may convert it into net gain - and the development was triggered by the birds not reproducing so well. Furthermore, when it comes to breeding performance in the Cormorant, the interesting level is the colony, or even the local population, comprising some ten or twenty large colonies, and the dynamic response of this larger unit to the onslaught of interacting environmental influences. Long-lasting or gradually accentuated environmental factors like general protection/persecution or eutrophication/clarification of both lacustrine De Nie 1995 and marine habitats - are likely to create diverse dynamic, highly upsetting responses on the time-scale of a few generations (10 - 30 years); any significant change in the same factors again likely to tip the scales, all the time overruling or disguising the influence of some inherited and adaptive individual performance.

To make matters worse, the particular events bringing about breeding failure (or success) are not the same from area to area, and from time to time. Pollution has its centre and its outskirts, although more and more coinciding from day to day. Tree-breeding and ground-breeding colonies are not equally exposed to wind, rain and predation, and disease doesn't spread at the same rate. Different energy densities of the fish prey in different areas and at different times of year are likely to affect the energy balance of adults (Platteuw & Van Eerden 1995) as well as the growth rate of young (Platteuw et al. 1995) and their endurance to wind, cold and rain. So, there is a pedagogical gain in repeating measurements and comparing results from different areas over and over again; the particular factors up till now promoting a spectacular population growth are likely to be brought out in full relief by such an approach. In this repetition we intend to favour a dynamic approach in a recent-diachronical perspective, emphasizing the effects of contemporary "events" rather than the evolutionary perspective and the possible guiding signals from the genetic level.

2. Material and methods.

Cormorants Phalacrocorax carbo sinensis have bred on two islets, Eskilstorps holmar (55° 29' N, 12° 56' E) in SW Scania, Sweden since the early 90s. A few nests were built here in 1991 and 113 pairs could be counted in 1992 (Fåglar i Skåne = "Birds in Scania" 1991, 1992). The islets are situated some 1.4 km from the mainland shoreline to the east, their present area is some 8 hectares. The surrounding water bodies could be called a Scanian "wadden sea", with 0.2 - 0.5 m water depth between mainland and islets, to the west and south there are protected nature reserves with grazed meadows. In the early 20th century both islets had thriving Arctic tern colonies and diverse wader communities, for a period in the late 1990's Herring Gulls Larus argentatus and Great Black-backed Gulls Larus marinus were the dominant species, gulls probably gradually deserting another colony on islet Måkläppen, 55° 21' N, 12° 49' E, after heavy fox, marten and mink predation reduced its breeding success to almost zero in 1990 (Fåglar i Skåne 1990).

For fourteen years, between 1991 and 2005, the dominant community of Eskilstorps holmar was this strange and somewhat appalling ground-breeding Cormorant colony (order of magnitude before it expired: 500 - 1000 pairs). The gulls bred in between, but many brought their young to the very borders of the Cormorant nests, or between them, establishing themselves both as refuse collectors and kleptoparasites, in 2005 as outright predators. As far as we were able to record they did little traditional food-collecting (a lot of strawberry kernels in Herring Gull pellets in 2005) on their own, and as a matter of fact most of the early gull broods in 2005 were eaten by gulls. Between 50 and 100 Eider Ducks Somateria mollissima also lay eggs on the islets each year; most of these broods are predated by the gulls. Eider Ducks that do not leave their nest for a second in the daytime will hatch, but many of their young still get predated on adjacent waters. In addition half a dozen pairs of Barnacle Geese Branta leucopsis laid eggs and hatched them in 2005 - probably due to active male guarding of both nests and females. In 2006 two dozen goose pairs lost all broods to predation. An odd occurrence is 1 - 2 pairs of rather aggressive Caspian Terns Hydroprogne caspia (replacement broods hatching in both 2004 and 2005, one nest predated in 2006). Nests or young of Arctic Tern Sterna paradisaea tended to get trampled down by the Cormorant creche; two young fledged in 2004, nothing in 2005, late replacements by 40 - 50 pairs in 2006 produced a dozen fledged young.
In general it can be said that a state of emergency, with zero or next-to-zero reproduction, rules in all nature reserves in SW Scania, due to the ongoing, extreme urbanization of the surrounding areas. Our "breeding" birds probably all originate from the Isle of Saltholm in the Sound (Öresund), and this situation has prevailed for a long time now; Eskilstorps holmar is one of the few, local enclaves where there is still some reproduction in some years. A lack of management is responsible for the situation; all reserves in the Öresund area should be brought together in a loose national park system with a central, responsible and informed management, such a measure would vastly improve the state of our egg-laying and to a great extent: egg & young-losing bird populations. It is our general opinion that the activities performed in the same area by The Scanian Ornithological Society and Falsterbo Bird Observatory haven't benefited the breeding bird populations for the last two or three decades; the "ornithological" presence has served mainly to legitimize the consecutive decline of breeding bird populations.

Human visits to the Eskilstorps colony in the daytime will put eggs and young of species with exposed nests at risk; when feeding their own young, gulls are likely to predate anything vulnerable and unguarded. Later on in the season enormous Cormorant creches used to roam the islets, trampling whatever came in their way and viciously biting back at attacking gulls. The general state of the overall bird community at best was a sort of armed truce, where species communicated with pecking bills. Knowing this, and still wanting to collect biometry elucidating the energy "balance-sheet" of this community we asked ourselves: how can biometric data be collected without upsetting the fragile balance of power, without patronizing or disturbing either species? Since we had some prior experience of nocturnal catching of waders, we decided to wait until the Cormorant young were well grown and to visit the colony mainly during dark hours, ringing and collecting biometry by the light of forehead torches. Early broods were not counted, and there were no visits to the colony in April or before 25 May.

The fishing-grounds of the adult birds lie 4 kms to the W, where Flintrännan meets Kogrundsrännorna ("gullies" in the Sound, > 1.0 % salinity) and onwards to Lillgrund, north of Skanör (the selected site for a huge, extremely wrong-placed windmill farm), furthermore at Stenudden, Ljunghusen (the Baltic; < 1.0 % salinity), 10 kms to the south and onwards, and south of islet Måkläppen (the Baltic), 15 kms to the SSW and onwards. These are shallow waters, the depth barely exceeding 5 metres where fishing Cormorants can be spotted from the shores. We have seen some discarded prey (Eel Anguilla anguilla and Eelpout Zoarces viviparus) in the nighttime, and young have regurgitated Eelpout. These were all rather large specimens; the major prey: flatfish Pleuronectiformes, sandeels Ammodytes sp. etc. probably dissolving too quickly to be recognizable when regurgitated. At any rate such items are removed by the gulls in the first morning hour, and we haven't collected anything for our own purposes, again minimum influence on the bird community has been our first objective. In 2005 the fishing flocks flying SW were smaller than in 2004, and large flocks flew E and NE for freshwater prey in Scanian inland lakes, probably lying as much as 30 km away. The issue of stress from variations in salinity on Öresund fish populations could be mentioned here, too; in general local populations are thought to be adapted to such variations, but we think that there is some emigration between water bodies connected with shifts in overall current directions, such shifts are likely to effect Cormorant fishing conditions as well, rippling of the surface not being the only problem connected with wind.

The main advantage with nocturnal ringing is the low level of disturbance and the torpidity of the young. When flushed, adults will alight fifty steps away and wait there, broods will remain collected in the nests. Gull parents will return to their young rather soon, and the two other Cormorant subcolonies remain undisturbed; we kneel on the ground and our activities are screened by luxurious vegetation. Under such circumstances we have ringed 805 juveniles and collected a wealth of biometry; growing wings measured straightened and flattened with a rule, weights taken with a spring scale with error 10-15 g in 2003, with an electronic scale to the nearest 5 g in 2004 and 2005. 27 young were measured and weighed twice in 2004, 33 in 2005. The June weather 2003 - 2005 is summarised in Table I, winds were some 50 % force Beaufort 2-3 and some 40 % Beaufort 4-5 in all three years:


Table I. Mean temperature, total precipitation and number of sun-hours in June 2003 - 2005, Sealand, Denmark. Data from DMI.



YearMean temp.
(deg. C)
Precipitation
(mm)
Sun-hours

200316.043246
200413.790173
200515.252239
Average15.152218



The weather was less summerlike in 2004 than in the other years (in particular during the ringing-period): June 2004 saw 17 rain-days against 11 in 2003 and 12 in 2005 and the mean temperature was below average, but the differences in wind force were only marginal. No ringed juveniles were found dead in 2003 or 2004, but we do not exclude mortality unseen by us; an infection with symptoms reminding of Newcastle disease took some small toll among both gulls and cormorants both years. In 2005 we saw no signs of disease (only starving young dying on nests), but predation in combination with human disturbance in the daytime took some toll; some 60 young were pulled from nests and had their stomachs opened by gulls, and hundreds of eggs must have been taken. So, with three years of fieldwork there is already a variety of backgrounds against which to analyse the material. Introduction to the complex of Cormorant problems has been our main purpose, therefore the opening question to the material has been: how do the weights of young develop while they are still confined to the nests or their immediate vicinity? And do the results conform with results derived from e.g tree-breeding colonies, where birds prey mainly on freshwater fish? For a preliminary evaluation we have calculated data from our six visits to the colony in 2004: 7, 9, 13, 17, 20 June and 3 July, and the nine visits in 2005: 26, 30 May, 8, 20, 21, 23, 26, 30 June, 10 July; these results are presented here.

young cormorant


Ill. 1. A member of Pelecaniformes! The young are a little torpid when ringed in darkness; broods will huddle together or act dead in the nests. Half-grown nestlings look like poodles; we well understand why authorities don't recommend clubbing young Cormorants with baseball bats, the anonymous spraying of eggs with mineral oil is less importunate to sensitive minds. Eskilstorps holmar 11.6.04.


2.1 Note on level of synchronization and predation: On 7 June 2004 some young were already fledged, which came as a surprise to us, since the colony development was less advanced by the same time in 2003. These early fledglings must have hatched by mid-April, the eggs laid no later than mid-March. Later on, in June 2004, there was a period of poor foraging, connected with strong winds. On 26 May 2005 a score of birds was again almost fledged, but the overall colony was retarded and much more fragmentised than in the two earlier years. One cause for this must have been the recurrent floodings of the islets in the preceding autumn and winter; many nests had been damaged and there seemed to be a general lack of nest material, nests were not as elevated as in 2004, some broods lying almost on bare ground. Furthermore the synchronisation of the breeding onset deteriorated gradually over the period 2003 - 2004 - 2005. (We do not think that we know all causes for this falling-off yet). A pre-hatching subcolony (southern islet) on 30 May 2005 had 2x1, 11x2, 14x3, 13x4, 2x5 eggs, average clutch-size ca ten days prior to hatching 3.10 eggs, subcolonies with young just hatching on the northern islet on 9 June had 9x1, 13x2, 41x3, 25x4 and 2x5 [eggs + young], average clutch-size 2.98 eggs. A new feature in 2005 was predation; Herring Gulls Larus argentatus and Great Black-backed Gulls Larus marinus pulling out young ca 1 kg of weight from the nests, opening their stomachs and devouring the contents, we counted at least 60 carcasses of young killed in that way, but only three ring-birds were found. Most of the predation took place between two of our own visits, we noted that the Caspian Tern young had been ringed by an alien ringer, an outright sabotage of our own work. (Only subcolonies in a particular development stage are vulnerable; the very energetic self-defence of older young is their salvation). In general the gulls themselves seemed to lose their own first laying to conspecifics in 2005, early predation being almost 100 % in all species except Cormorant, and it was not until well after midsummer that there was a fair amount of newly-hatched gull chicks. In our opinion both gulls and cormorants experienced some degree of food scarcity in May and early June 2005, the Cormorants compensating by partly resorting to distant freshwater foraging (benefiting from an expansion of their niche), while the gulls switched to overt predation instead of the kleptoparasitism that characterised the balance of power between gulls and cormorants in 2004. In 2006 - with Cormorants absent - the mortality among Gull chicks was again surprisingly high; Great Black-backed Gulls raising practically nothing and many Herring Gulls being predated (but left uneaten) just prior to fledging.

3. Results

3.1. The overall material

Wing lengths of young ringed in nests or caught in the creche range between 60 - 70 mm (feather in sheath, but vane soon emerging) and 290 mm, weights between some 550 and 3,170 g. A nestling with wing-length 70 mm, weighing 700 g is c12 days old at Eskilstorp, with wing-length 250+ mm and weight 2000+ g the birds run for the water when disturbed - to a great extent avoiding being caught. By then they should be some 38 - 40 days of age. Hence there is a period of approximately one month when a juvenile can be handled and examined; before that a ring with inner diameter 16.5 mm can be pulled off, after that the birds evade the ringer. The ratio [weight : 10 x wing] has been calculated for each bird and the material divided into 7 groups, the results for 2004 and2005 are presented in Table II:


Table II. Ratio weight : 10 x wing in young Cormorants for eight wing-length intervals, 2004 and 2005



wing-length
(mm)
n (2004)weight : 10 x wing
± 1 s.e.; 1 s.d.
n (2005)weight : 10 x wing
± 1 s.e.; 1 s.d.

50 - 80--550.98 ± 0.02; 0.13
81 - 110161.01 ± 0.03; 0.12570.99 ± 0.01; 0.09
111 - 140321.00 ± 0.02; 0.10350.97 ± 0.02; 0.11
141 - 170401.00 ± 0.01; 0.10370.95 ± 0.01; 0.09
171 - 200520.96 ± 0.01; 0.11210.94 ± 0.02; 0.09
201 - 230410.96 ± 0.01; 0.08390.90 ± 0.01; 0.08
231 - 260240.90 ± 0.02; 0.11380.89 ± 0.02; 0.13
>26060.90 ± 0.03; 0.0670.81 ± 0.03; 0.09



The same materials are shown with weight plotted against wing-length and linear and quadratic regressions drawn in Fig. 1 and Fig. 2. Both variables are Normally distributed, of different dimensions, with random measuring errors, and should be treated as a Model II regression (Sokal and Rohlf 1995), with focus on correlation and coefficient of determination. The equations of the regressions are:

2004:

WEIGHT = 179.13 + 8.59 WING; Radj2 = 82.3 %; F = 956.4, P = 0.000

WEIGHT = -125.4 + 12.26 WING - 0.0103 WING2; Radj2 = 82.6 %; F = 487.4, P = 0.000

2005:

WEIGHT = 140.01 + 8.36 WING; Radj2 = 90.9 %; F = 2868.2, P = 0.000

WEIGHT = -116.56 + 12.26 WING - 0.0124 WING2; Radj2 = 91.4 %; F = 1524.2, P = 0.000

On the whole, the figures indicate approximate isometric growth, the mass growth rate being a constant function of body size as measured by the wing length, 82 % (2004) and 91 % (2005) of the weight variation explained by variation in body size as measured by wing length. The assymptote that puts an end to growth barely asserts itself at the end of the interval, and still the material from 2005 covers at least 30 days of nestling growth. It is obvious from both Table II and Figs. 1 and 2, however, that neither weight : wing ratio nor growth rates are constant up to the horisontal slope of the real asymptote; between weights 2.0 and 3.0 kg the ideal isometric relations gradually recede, a fact that is also evident from the values of Table II. In order to investigate possible influences from adverse weather, the [weight : wing] ratio was plotted against time and the overall regressions calculated, the equations are:

2004:

RATIO Weight/Wing = 1.07 - 0.0079 DATE Radj2 = 10.8 %; F = 25.63, P = 0.000

2005:

RATIO Weight/Wing = 0.94 +0.0002 DATE Radj2 = 0.0 %; F = 0.15, P = 0.697

The material is shown in Fig. 3 and Fig. 4; the very horisontal regression from 2005 (which was by no means a breeding season without problems; there may have been adequate weight, but we suspect that the quality of the prey wasn't as good as in 2004) indicates that the slight decline of the ratio in 2004 reflects some real adversity to the birds, most likely connected with wind and poor visibility in the water (the whole water body had been much silted for a sequence of preceding years due to construction works connected with the Öresund bridge). The lower coefficient of determination in the 2004 material also speaks in favour of this view.

3.2. Brood size, development within broods, effect of ranking

We ringed birds from 113 discernible broods in 2003, 98 in 2004, 144 in 2005; more birds leaving nests and joining creches in 2004 and 2005 than in 2003, a few broods may have been taxed too low because of one large young escaping. In addition predation gradually reduced broods in a couple of subcolonies in 2005. If broods were visited and handled twice we have used the earliest value for brood size. Runaways still return to nests, sitting with their kin on the nest two or three days later, so the brood size has also been corrected upwards in a couple of cases. The mean number of hatched and live young sitting on nests (age 20 - 30 days in 2003 and 2004, 12 - 30 days in 2005) per recognized brood was 1.94 ± 0.08, s.d. 0.85 in 2003, 2.18 ± 0.08, s.d. 0.82 in 2004, 2.17 ± 0.08, s.d. 0.95 in 2005 (normal statistics), the material is presented in detail in Tables III and IV. The true 2005 value is lower than the one offered; we were unable to connect predated young with broods. The total number of unhatched eggs from 211 broods of course was much higher than 42; unguarded eggs will be taken by the gulls whenever there is an opportunity.


Table III. Number of broods with 1 - 4 young and the number of unhatched eggs at Eskilstorps holmar in June-July 2003-2005.



Yearnestlings aliven1 unh. egg2 unh. eggs

200314053
20032443-
20033252-
200344--

2004123--
200423631
2004337--
200442--

20051389-
200525881
20053342-
2005413--
200551--




Table IV. Broods with 1 - 2 and 3 - 4 live young and overall percentage of broods with 1 - 2 nestlings, Eskilstorps holmar 2003-2005.



Year1 or 2 young3 or 4 young% 1+2, 1 binom. s.e.

2003842974.3; 4.1
2004593960.2; 5.0
2005964866.7; 3.9


Both np and n(100-p) > 500, so the difference between percentages may be tested with normal distribution procedures; Z = 2.18, i.e. the difference between 2003 and 2004 is significant at the 5 % level. Similarly a G-test of independence gives adjusted G = 4.76; the null hypothesis that both years had the same ratio of small and large broods can be rejected at the 5 % confidence level. No other differences are significant.

Within 2- and 3-broods young can always be ranked according to size. In 2004 we took measurements from 34 three-broods, 30 two-broods and 18 one-broods, in 2005 from 34 three-broods, 61 two-broods and 36 one-broods; mean values for these data have been brought together in Table V. Predation losses hamper the 2005 material, and a few values from three-broods were rejected for writing or measuring errors. The variances of several categories exclude the use of ANOVA.


Table V. Weight : wing ratio and mean wing length for young from broods with 1, 2 or 3 young, 2004 and 2005. The largest young was ranked 1, etc. The 2005 value for the smallest young of 3-broods was discarded because it was generally taken at a later stage than the values for young 1 and 2.



Brood size;
rank of young
nweight : 10xwing
± 1 s.e.; 1 s.d.
mean wing length
± 1 s.e.; 1 s.d. (mm)
2004
3; 1331.00 ± 0.02; 0.11192 ± 6; 36
3; 2340.97 ± 0.02; 0.09186 ± 6; 36
3; 3310.94 ± 0.03; 0.15161 ± 8; 43

2; 1300.99 ± 0.02; 0.12175 ± 8; 45
2; 2300.95 ± 0.02; 0.10151 ± 8; 45

1180.98 ± 0.02; 0.10163 ± 11; 46
2005
3; 1340.98 ± 0.01; 0.10134 ± 11; 63
3; 2260.96 ± 0.03; 0.15123 ± 12; 61
3; 3140.95 ± 0.03; 0.11discarded

2; 1610.95 ± 0.01; 0.10151 ± 8; 63
2; 2530.95 ± 0.01; 0.10141 ± 8; 53

1360.94 ± 0.02; 0.12148 ± 10; 60


3.3. Individual development, daily growth rate, asymptote

Data for birds measured twice in 2004-2005 are summarised in Table VI:

Table VI. Growth rate of wing and weight in birds measured twice during the nestling period, 2004 - 2005.



YearnMean growth-rate of wing-length
± 1 s.e.; 1 s.d. (mm/day)
Mean growth-rate of body weight
± 1 s.e.; 1 s.d. (g/day)

2004279.3 ± 0.4; 2.260.0 ± 6.0; 30.8
20053310.3 ± 0.3; 1.697.1 ± 4.4; 24.9

We cannot exclude the possibility that there has been some bias in our selection of birds, affecting the recorded weight increase, but at least the two materials have similar variances (F = 1.53; the critical value at P=0.05 is 1.83) and a z test shows that the daily weight increase in birds weighed twice was significantly higher in 2005 than in 2004 (z = 5,05, p<0.01). This in turn affected the mean growth-rate of wings (z=1.97, p=0.05), but not by the same proportion, and even in birds barely increasing their weight, wings continued to grow by 5 - 6 mm/day. The highest growth-rate in 2004 was achieved by a bird increasing by 740 g (from 1540 g) over 6 days, growing its wing by 76 mm (12.7 mm/day) over the same interval, in 2005 one bird increased by 460 g (from 630 g) over 3 days, increasing its wing-length by 33 mm (11.0 mm/day) over the same interval. Single birds have sat dying on nests in all three years, sick or for some reason deserted by the parents, the material may include a few measurements from such birds, but they barely affect the calculated values, and we have chosen to discard as little material as possible.

From Figs. 1 and 2 the pronounced isometric relation weight : wing is evident, at the same time quadratic regressions are slightly better than linear ones both years; there is some small decline over time in the growth rate. This calls to mind the comment by Ricklefs 1967:"...the Gompertz, and to a greater extent the von Bertalanffy curves differ from the logistic curve in a marked slowing of growth rate and prolongation of the growth period in relatively later stages of growth." Referring to this observation we have chosen the Bertalanffy equation for our model; with data from biometric repeats it enables us to calculate approximate values of an overall asymptotic weight M and of the growth rate K. In Figs. 5 and 6 we have plotted the third root of [weight at ringing + 1 daily rate of weight increase] versus the third root of weight at ringing, the intersection with x = y gives the third root of the asymptotic weight, or asymptotic weights 2300 and 2365 g in the two years. Assuming that the line is an estimate of e-K, we get growth rate K = 0.080 and 0.106/day and k, the rate of destruction of mass per unit mass = K/3 0.027 - 0.035/day.

Finally, the pooled repeat materials from 2004 and 2005 serve as a basis for three instructive regressions: Fig. 7, Fig. 8 and Fig. 9; equations for the lines are:

Wing growth rate = 6.81 + 0.0388 Weight growth rate Radj2 = 46.8 %; F = 51.18, P = 0.000

Weight growth rate = 145.7 - 0.056 First weight Radj2 = 48.1 %; F = 53.82, P = 0.000

Wing growth rate = 12.98 - 0.027 First wing length Radj2 = 33.4 %; F = 30.65, P = 0.000



eskilstorps holmar


Ill. 2. The original relation between gulls and cormorants in the Sound area may not only be "commensalist", but even symbiotic, there is no fish waste around the nests, and some adult gulls appear as keen kleptoparasites when Cormorant young are fed. On the picture we have withdrawn twenty steps from two creches; gulls are already closing in, and two have landed. We have nothing but circumstantial evidence so far, but it seems to us that "the parental fishing effort" of Cormorants in this colony involves some effort on behalf of gull chicks as well. This would probably not be possible without prey of high energy density. As an extra bonus, the marine fishing-grounds are not particularly distant: their proximate limits 4 - 15 kms away. Eskilstorps holmar 11.6.04.


4. Discussion

Clutch size, brood size at hatching, fledging success, unhatched eggs.

1. Breeding success similar to that in neighbouring areas. Van Eerden & Gregersen 1995 give values for breeding success in various colonies in Denmark, Germany and The Netherlands, the period covered is 1980 - 1990. Data refer to the average number of young per nest that reach the age of 25 - 35 days (which is when the Eskilstorp birds are ready to leave the nests for the creche). In good years the large colonies in Denmark produce more than 3 young per succesful nest, if the failed broods are included as well, the mean value is 1.9 - 2.4 young per nest, values much the same in Germany, while the lowest values in The Netherlands lie well below 1 young per nest. If we take the mean clutch in all broods (total material 355 broods) where birds were ringed: 1.94 (2003), 2.18 (2004) and 2.17 (2005), and assume that all birds ringed got fledged - and they were in 2003 and 2004 - the breeding success of a ground-breeding colony at wind-exposed Eskilstorps holmar is much the same as that of tree-breeding colonies in Denmark and Germany. In a Polish colony Stempniewicz et al. 2000 lost 1/4 of the production during a period of strong winds in June 1996, still the fledging success only fell to 2.19 young per nest against 2.45 the year before. This could be compared with a mean of 3.00 eggs in 132 broods from Eskilstorps a week before or just prior to hatching in 2005. Here we wish to draw the attention to the number of unhatched eggs in broods where young were already large enough to be ringed, it is striking, and the real number is much higher than the 42 seen by us, it should be numbered by hundreds. These eggs should be investigated for their contents of different pollutants; we mustn't forget that the Sound drains the critically polluted Baltic, and the outlet of the major sewage treatment plant of the polluting Malmö region is situated only a few kilometers to the north of Eskilstorps holmar.

Asymptotic weight and growth rate, effect of ranking within brood

1. Asymptotic weight dynamically "fine tuned". There is a period of almost perfect and rather steep isometric growth in many species where young are fed on high-caloric fish prey, e.g. Little Tern Sterna albifrons (Norman 1992), Common Tern Sterna hirundo and Sandwich Tern Sterna sandvicensis (Klaassen et al. 1992). Platteuw et al. 1995 showed that the growth curve of nestling Cormorants from The Netherlands is of the same basic character (but more detailed diagrams in the Dutch paper might have allowed more exhaustive, and possibly slightly deviating, conclusions). In both Pygmy and Great cormorants Shmoely (2001) observed that the growth rate constant (K) was higher than predicted from the allometric equation [y = xa; our comment], based on the asymptotic body mass of the chicks. A possible reason for this may have been a too low estimate of the asymptote (due to model and equation) - or that environmental constraints de facto create "premature" asymptotic weights, or that asymptotic weights are dynamically determined, changing from case to case and from year to year (with food amounts, food quality and sex ratio). Our own data indicate a dynamic fine tuning (Figs. 5 and 6) of the asymptotic weight, it is dynamically "aimed at" at an early stage of development, and the general character of this dynamics is that of a linear attractor. Platteuw et al. 1995 had a wide range - wider than expected - of asymptotic weights in their sinensis colony (1781 - 2440 g), but in six birds from 3 broods with 3, 2 and 1 young, the final weights did not exceed 2000 g in either case, prompted with this comment by the authors: the last recorded body mass among these nestlings provided only a slight underestimate of the best estimate for asymptotic body mass. Again, these asymptotes may have been "artefacts", created by the young being prematurely unavailable for weighing.

2. The distribution of the asymptotic weights is bimodal. From the Bertalanffy equation and Figs. 5 and 6 we calculated provisional average values for asymptotic weight in 2004 and 2005: 2,365 and 2,300 g. Since sexes are not equally large, there will in practice be a bimodal normal distribution of asymptotes, our calculated values probably lying between a female average of, say, 2,200 g and a male of 2,700 g (Koffijberg & Van Eerden 1995, Engström 2001). With this observation at hand, we would be very interested to know if sex ratios in Cormorant colonies are affected by food availability, egg weights or outright manipulation.

3. Relative variance in weight larger than variance in wing-length. Our own Figs. 1 and 2 illustrate a good isometric relation weight (g) : 10 x wing (cm) = 1 over practically the whole growth interval, with many birds well over 2000 g, up till the moment when nestlings enter the water and in practice become unavailable for further weighing. On the other hand Table II bears evidence of a slowing-down of wing growth and a beginning asymptote to weight growth somewhere above wing-length 200 mm. Rank within clutches (Table V) is one reason for individual differences, but there is some contradiction between Table I and Table V on this point; if the ratio weight : wing decreases with age and size, the advantage of the largest young must be even more pronounced than indicated by the values of Table V. The ratio weight : wing decreased as the season progressed in 2004 (Fig. 3), but not in 2005 (Fig. 4); after comparing the figures and considering the different degrees of synchronization (better in 2004) we suggest that at least part of this decline was caused by windy weather and disadvantageous fishing conditions, the difference between young within broods maybe accentuated in 2004 with lasting adversity. Part of the decline may simply be seasonally conditioned, however, by the fact that all young gradually approach the asymptote; with development of muscles and competitive and defensive behaviour the energetic expenditure for movement increases. If we turn to the extremes, some deflections in juvenile growth are probably caused by differences in parental experience and feeding effort, Platteuw et al. 1995. The young of some 3- and 4-broods are always fed to the point of saturation, at the other end of the scale a few broods seem to be on the verge of starvation. In the end, stored body reserves will act as a buffer, guaranteeing some minimum and constant rate of remige growth; the standard deviation of the weight growth rate was more than 50 % but only 25 % of the wing growth rate (Table VI).

4. Weight : wing ratio higher in the early stages, positive correlation between mass and size growth rates. Figs. 8 and 9 both show, that the growth rates for body mass and wing-length are highest when young approach the "ringing size", with wing-lengths in excess of 50 mm and weights above 500 g. From this point onwards there is a delicate weighing-up between food intake, available metabolic capacity, catabolic build-up and anabolic losses, that keeps the ratio between body mass and body size as measured by wing-length almost constant. In this context the applicability of the Bertalanffy model to Cormorant growth is almost wondrous, and there is no clumping of data at either end. These two figures in turn demonstrate that there are constraints to the positive correlation between mass growth rate and wing-length growth rate illustrated by Fig. 7: the highest values occur when the nestlings are 10 - 15 days of age, provided that they are adequately fed.

5. Should a "double" Bertalanffy growth curve be applied? Cormorants may seem a little non-conforming to theory in some of the contexts mentioned above, but the fault need not be with the theory; sometimes the blame is with assumptions underlying calculations, not only assumptions regarding the "asymptote" M - when calculated, the value to some extent depends on the choice of growth equation - but also assumptions about some constant caloric value of fish prey and its reflection in individual (inds. belonging to a particular colony, area) growth-rates and asymptotic weights. Judging from our own material and from what we have been able to extract from literature, we expect the "asymptote" to show some minute slope for maybe a month after it has come into effect (indicating that a "double" von Bertalanffy growth curve should be used if the biometric material is extended over, say, 2 months - cf. Cloern & Nichols 1978, Condrey et al. 1988, Vaughan and Helser, 1990 - and we further expect this value, both the slope up to the first turning point and the slope of the following "asymptote", to be influenced by the energy density of food brought by the parents.

Energy density of fish prey

1. General energy density in fish. National alimentary databases, e.g. Livmedelsverkets livsmedelsdatabas and USDA National Nutrient Database for Standard Reference, supply values of body composition and caloric content for a wide range of fish consumed by man. There is a fault in these estimates, however; in most cases they give the energy content of "portions", filets etc. In fish species with low average fat content or none at all, standard values are 80 % water and 16 - 20 % protein (cod, cyprinids, percids), this gives energy values of some 350 kJ (85 kcal) per 100 g fresh weight. Not considered then are the energy contents of liver and gonads, which contain more fat than body tissue in many species, in particular prior to spawning. Some marine species store large amounts of fat in body tissue as well; variations in energy density between species and seasons are primarily connected with fat content in such species. So, when Klaassen et al. 1992 adopt an overall energy density value of 5.0 kJ/g fresh weight for Cormorant prey, the value based on Pikeperch Stizostedion lucioperca data from August and September presented in Buijse & Houthuijsen 1992, there is reason to check the accounts; these authors are generalizing values from a rather lean percid to be valid for the energy density of all fish consumed by Dutch Cormorants during the breeding season.

The Buijse & Houthuijsen values are 19.8 - 22.5 kJ/g dry weight, resulting from bomb calorimetry and presumed gross energy values 23.5 kJ/g protein and 39.5 kJ/g fat. Recalculating their values, we observed that the Pikeperch contained 80 % water, 17 - 19 % protein and 0 - 3 % fat, in addition there should be some 1.25 % of ash. With a generous estimate this is equivalent to 0.18 x 23.5 + 0.025 x 39.5 = 5.22 kJ/g fresh weight. Fat is more digestible than protein, and the metabolization losses are higher in protein than in fat; for salmonids Brett & Groves 1979 suggested metabolizable energies of dietary fat 8.0 kcal/g and of dietary protein 4.2 kcal/g, which would reduce the energy density to 4.00 kJ/g. In addition fish lose less energy in urine (uric acid) and have less heat increment, we believe an assimilation factor 0.75 (which is the rate of energy assimilation from optimal food in mice; e.g. Koleja et al. 2001: J. Exp. Biol. 204: 1177 - 1190) to be more accurate in birds, corresponding to metabolizable energy 3.90 kJ/g fresh fish mass. And after spawning in April - May, coinciding with maximum food demand from Cormorant nestlings, the energy content will be at its minimum in mature percids and cyprinids.

2. Fat fish, preferably marine species, have higher energy densities. So, during the part of the breeding season with highest energy demand, fairly lean percids and cyprinids make up most of the food intake of freshwater-fishing Cormorants in NW and Central Europe. Their hegemony does not extend to the brackwater Baltic or the Sound/Belt area, however, here marine species are added to the diet list and other energy densities may become possible. In freshwater Pikeperch the lipid content grew with age (Buijse & Houthuijzen 1992), but this state of things is likely to get inverted in marine habitats; in 15 species from sub-Antarctic waters Tierney et al. 2002 found that the smallest size classes generally had the highest caloric content, with energy densities ranging between 22 and 59 kJ/g dry weight. Sandeels Ammodytes sp., important prey to many seabirds, seem to have only average energy densities, the same as gadoids in the North Sea area; we calculated values 3.8 kJ/g fresh weight with 20 % dry mass, 5.5 kJ/g fresh weight with 25 % dry mass from a formula offered by Pedersen & Hislop 2001. Baillie & Jones 2003 found energy densities 6.0 and 6.4 kJ/g fresh weight in Sandeels from the Labrador area. Robards et al. 1999 concluded: The seasonal food value of adult sand lance to predators varies markedly, but maximum energetic value coincides with important feeding periods for marine mammals, fish and seabirds. It could be added, that there are two groups of Sandeels in the Baltic; one spawning in spring, one in the early autumn. The difference between 4 kJ/g or less and 6 kJ/g or more may seem insignificant, but really fat prey of marine origin may double the energy content offered by the poorest cyprinids after spawning.

Schooling marine species like Baltic Herring Clupea harengus have lipid values ranging from below 10 in summer to over 20 % in winter, added to some 18 % protein and 60 - 70 % water. In the former case, with the lower, "metabolizable" conversion factors the energy density would be appr. 6.5 kJ/g fresh weight, in the latter 10 kJ/g or more. Another important prey species has not been mentioned: the Eelpout, Zoarces viviparus, in the past major prey of Baltic and Belt Sea Cormorants, but largely ignored in investigations of prey preference and depletion (Jonsson 1979, Hald-Mortensen 1994, Engström 2001). The viviparous Eelpout "spawns" late in summer, in August and September, and will be at its highest energy density by that time. We haven't found any relevant information concerning the energy density of this species, but Förlin 1999 found gonad weights 11.5 - 19.4 % of total body mass in Bohuslän, W. Sweden, in autumn. N. Jepsen, Danish Inst. for Fisheries Management (in litt.) produced the following educated guess: the Eelpout probably lies around 700 Kj/100g in the period before spawning. Our point here is: with some effort on part of the birds, or at an advantageous site, Cormorants may achieve energy densities in prey of maybe twice the minimum value. A wider range of bomb calorimetry results, involving all possible prey species, would help to assess the advantages and bottlenecks to Cormorants; for example, if young hatch in April they will most certainly be fed high-caloric pre-spawning percids or cyprinids, and if adults exploit Eelpout populations in June and July, these will probably be at their maximum energy density by that time. We see no other plausible explanation for the fact that birds weighed twice in 2005 had 60 % higher daily weight increase, while the wing-length growth rates differed by only 10 %.

  • Energetik: referenser, A - K, abstracts och kommentar. / Energetics: references, A - K, abstracts and comments.
  • Energetik: referenser, L - Z, abstracts och kommentar. / Energetics: references, L - Z, abstracts and comments.
  • Skarvreferenser, författarnamn A - M. / Cormorant references, author's name A - M.
  • Skarvreferenser, författarnamn N - Z. / Cormorant references, author's name N - Z.
  • Till kärrsnäppesidan/To the Dunlin Web Site
  • Back to start page


    Published 26.10.05, last corrected 16.1.07.