Other formats

    Adobe Portable Document Format file (facsimile images)   TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Proceedings of the First Symposium on Marsupials in New Zealand

Results

Results

Seasonal distribution of births

In the Orongorongo Valley most births occur during April-June (Fig. 2) as noted by Kean (1971, 1975) and Crawley (1973), the mean date of birth for 494 pouch-young over 1966-75 being 15 May (median 18 May). Only 3 births (0.6%) were recorded in spring (i.e. after 1 September). During 1971-74 mean birth dates in areas A and B were 15 and 13 May respectively, and were not significantly different.

In both study areas females raising pouch-young had earlier mean birth dates than those which lost their young. However while the differences were significant in area A (dates 13 May and 1 June, t = 6.451, p<0.001), they were not so in area B (dates 6 and 14 May).

Mean birth dates for yearling and 2 year-old females were later than for those 3 years old and over.

page 95
Number of young

Females gave birth to only single young and no twins were recorded; Kean (1971) found only one set of evident twins during 1946-61 out of a sample of 8,000 Orongorongo females. In October-November 1972 two females carried fostered young in their pouches in addition to their own progeny (Bell 1974); all four young were furred, weights of the progeny being 220 g and 258 g compared with corresponding weights of 150 g and 230 g for fostered young. No cases of double-breeders, as reported for other areas (Kean 1971), were detected during 1966-75; but one was recorded in area A in 1978 (R.E. Brockie, pers. comm.).

Breeding success
Overall trends

In general most females gave birth to young each year (Table 1), and the birth rate over 1971-74 was not significantly different between the two areas (X1 2 = 3.1810, 0.10>p>0.05). Many young died in the pouch. Over 1971-74 the survival rate in area B (42%) was significantly lower than in area A (58%) (X2 1 = 4.1489, p<0.05).

The timing of the disappearance of pouch-young was estimated very approximately as the mid-point between their last sighting and the date the pouch was first found prematurely empty; mean dates were 7 September (area A) and 17 August (area B), while corresponding mean ages of young (±S.E.) were 95.4 ± 3.5 days (n = 107) and 97.9 ± 7.7 days (n = 33).

The estimated rates at which young surviving the pouch-phase were produced (i.e. birth rate × survival rate) were only 38% and 33% for areas A and B respectively. So although most females had young, their final productivity was low compared with some other reports (e.g. Dunnet 1964, Kean 1971).

Grey-coloured females predominated over blacks in area A (60:40) but not in area B (46:54); the difference between areas was significant (X2 1 = 5.8615, p<0.05). There were no significant differences in the measures of breeding success of the two colour morphs (X2 1 0.000 to 0.5109).

Age and Breeding success

The ages at which females first produced young were determined for 41 animals of known age and breeding history in area A:

page 96
Fig. 2 Seasonal distribution of possum births in the two study areas. The date of birth is scaled as 10-day intervals from January 1st with the months of May and September indicated by stippling.

Fig. 2 Seasonal distribution of possum births in the two study areas. The date of birth is scaled as 10-day intervals from January 1st with the months of May and September indicated by stippling.

page 97

7 (17.1%) first bred as yearlings;

23 (56.1%) at two years;

9 (22.0%) at three years;

and 2 (4.9%) at four years.

The mean age of first breeding was 2.15 years with no significant difference between blacks and greys (X2 3 = 1.3274. p>0.10).

Data from a larger sample (Table 1) show for area A over 1966-75 only 8% of yearling females gave birth, and none reared pouch-young successfully; no area B females gave birth as yearlings.

Approximately 60–65% of two-year-old females had young, birth rates between areas differing significantly over 1971-74 (X2 1 = 4.222, p<0.05). None of their young survived in area B, however, but 47% did in area A during 1971-74 (and 52% over 1966-75). Thus while no two-year-old females were known to rear young in area B, 34% did so over the same period in area A (Table 1).

The birth rate was significantly greater in adults than two-year-old females in both area A (X2 1 = 11.6710, p<0.001) and area B (X2 1 = 9.8776, p<0.01); it was also greater in adults in area B (93%) compared with area A (81%) during 1971-74 (X2 1 = 8.3811, p<0.01).

The 'adult' category (Table 1) comprised all animals three or more years old. Data for a more limited sample of precisely-aged possums are given in Table 2; sample sizes for older age classes (<5 years) are unfortunately limited. As seen in Table 1 the birth rate increases dramatically between yearling and two-year-old possums (X2 1 = 48.8057, p<0.001). Significant increases in birth rate also occurred between years 2 and 3 (X2 1 = 5.3171, p<0.05) but thereafter there were no significant differences between successive years except years 5 and 6 (X2 1 = 5.2179, p<0.05). An explanation, at least in part, for the lower birth rate of five-year-olds is evident when these age-specific breeding data are related to breeding success in different years of the study: most data on five-year-olds came from 1973 and 1975, two years when the birth rate was relatively low (see Fig. 3).

The survival rate of pouch-young increased significantly between yearlings and two-year-olds; the young of four-year-olds survived significantly better than those of either three or five-year-olds (X2 1 = 8.7273-9.3818, p<0.01). Over the first six years therefore, the highest female productivity occurred in four-year-olds, 91% producing young that survived the pouch-phase. The page 98
Table 1. Overall breeding success of female possums in relation to age, color and area.

Table 1. Overall breeding success of female possums in relation to age, color and area.

page 99
Table 2. Breeding success of known-age area A females in successive years of life.
AGE IN YEARS
1 2 3 4 5 6 7 8 9 10
BIRTH RATE % 8.2 63.6 85.0 90.6 71.9 100.0 91.7 75.0 100.0 100.0
N 85 55 40 32 32 15 12 4 5 2
SURVIVAL RATE OF POUCH-YOUNG % 0 52.0 63.6 100.0 63.6 66.7 25.0 50.0 100.0 0
N 5 25 22 21 11 9 8 2 1 2
PRODUCTIVITY RATE OF FEMALES % 0 32.8 54.1 90.6 45.7 66.7 22.9 35.5 100.0 0
perceptible decline in success after the fourth year is in part due to increasing numbers of poor-conditioned females in the samples. Those females known to have died in area A from natural causes showed a marked decline in breeding success in their last year (Table 3); the differences in birth rate and survival rate between the last two years of life were highly significant (X2 1 = 11.35 and 12.37 respectively, p<0.001); differences between other years were not significant. At death some females had lost their young, others still retained them; often their poor condition resulted in unclean, smelly pouches and conspicuously undernourished young.
Table 3. Breeding success of area A females up to and including the year of death.
YEAR DIED PREVIOUS YEARS
1 2 3
% N % N % N % N
BIRTH RATE 48.3 29 91.7 24 93.3 15 93.8 16
SURVIVAL RATE OF POUCH YOUNG 0 14 63.6 11 66.7 9 80.0 10
PRODUCTIVITY RATE OF FEMALES 0 58.3 62.2 75.0

Although slightly more one and two-year-old greys had young compared with blacks, no significant age-specific differences in breeding success were found between the two colour morphs in area A (Table 1; X2 1 =0.5241 and 0.2387 for yearlings and two-year-olds respectively, p>0.10).

page 100
Annual fluctuations in breeding success

There was marked variation in the onset and median dates of birth between years. The median date of birth in area A was especially early in 1971 (19 April); in other years median birth dates fell later as follows: 2 May (1966), 10 May (1969), 15 May (1970), 18 May (1974), 19 May (1967), 21 May (1972), 24 May (1973), 25 May (1968), and 4 June (1975). In area B the dates were 7 May (1971), 8 May (1972), 11 May (1974) and 15 May (1973). [unclear: These] annual fluctuations were also reflected in breeding trends of the sa[unclear: fe] female over consecutive years. Spring births (after 1 September) occurred only in area A and in 1967 (1) and 1971 (2).

Table 4. Breeding success in yearling and two-year-old females, 1966-75.
BIRTH RATE SURVIVAL RATE OF POUCH-YOUNG
(Proportion known to have pouch-young) (Proportion of breeders known to rear pouch-young)
YEARLING 2 YEAR OLD YEARLING 2 YEAR OLD
AREA A 1966 0/7 0/0 0/0 0/0
1967 0/10 6/9 0/0 2/5
1968 0/5 1/5 0/0 1/1
1969 2/12 0/1 0/2 0/0
1970 0/12 6/10 0/0 3/4
1971 5/16 8/8 0/3 3/4
1972 0/11 9/15 0/0 3/8
1973 0/5 3/4 0/0 1/2
1974 0/6 1/2 0/0 0/1
1975 0/1 1/1 0/0 0/0
AREA B 1971 0/6 1/1 0/0 0/1
1972 0/6 2/4 0/0 0/1
1973 0/6 0/1 0/0 0/0
1974 0/2 1/1 0/0 0/0
page 101
Fig. 3 Annual variation in breeding success of adult females in the two study areas.

Fig. 3 Annual variation in breeding success of adult females in the two study areas.

page 102

Among younger females, yearlings bred in area A in 1969 (2) and 1971 (5); two-year-olds bred in most years, but the birth rate was again high in 1971 (Table 4). Amongst adult females in area A there was significant heterogeneity in both the annual birth rate (X2 9 = 23.8998, p<0.01) and the survival rate of pouch-young (X2 9 = 37.2221, p<0.001), but not in area B over 1971-74 (X2 3 = 0.4005 and (X2 2 = 2.3723 respectively). These trends are illustrated in Fig. 3.

The approximate mean ages at which pouch-young were lost also varied, ranging from 74 days (area A, 1970) to 127 days (area A, 1966). In area A mean dates of loss fell between: (i) 16–31 August (1970, 1972, 1974, 1975); (ii) 1–15 September (1968, 1969, 1971, 1973) and (iii) 16 September -1 October (1966, 1967). Mean dates of disappearance of young ranged from 13–19 August over 1971-74 in area B.

In summary, there were productive years, such as 1971, when more females bred (including some yearlings), when births were earlier, and when survival of pouch-young was high; and less productive years when fewer females bred, births were later and few pouch-young survived. Correlations between these parameters are shown in Fig. 4.

Female weight and breeding success

Stored fat reserves in part reflect the general level of nutrition and well-being ('condition') of an animal. Thus, in female possums, such reserves could affect the ability to break anoestrous and have young, and to maintain such young while in the pouch. Since the study animals were live-trapped, direct estimates of fat-deposits were not feasible. Thus body weight (or a weight/length index) was used as an indication of fat reserves. By including and excluding weights of the young, both 'gross' and 'net' weights of females were determined.

Such a correction was more important when the young were older; the overall 3√weight/age regression (Y = 0.0365X + 0.7150) based on 264 young aged up to 250 days gave an estimated weight of 331 g at 170 days; different regressions were used for different years unless samples were small when this overall regression was used. Since young remain attached to the teat during early pouch-life, no weight data were available from the study areas; data from 52 young ranging from 0.7 g to 85.5 g (mean 29.1 g) from Waverley, N.Z. were therefore incorporated into the predictive growth curve; this was page 103
Fig. 4. Linear regressions and correlations between annual breeding parameters for adult females, area A (1966-75).

Fig. 4. Linear regressions and correlations between annual breeding parameters for adult females, area A (1966-75).

The proportion of adults breeding and the proportion rearing pouch-young are as defined in Fig. 3 but with arcsin transformation of the percentages.

page 104 considered justifiable since: (i) differences in growth rate between populations are less during early stages of development; and (ii) weights at such a stage are in any case a tiny fraction of the female's weight.
Overall mean weights of females

The overall mean weight for adult females in area A over 1966-75 was 2.32 kg (gross) and 2.27 kg (net), based on 2,191 captures of 207 individuals; during 1971-74 comparable means were 2.35 kg and 2.31 kg for area A (1960 captures, 108 individuals) and 2.33 kg and 2.20 kg for area B (606 captures, 89 individuals); mean weights for 1971-74 did not differ significantly between areas. However, adult females were on average longer in area B (78.7 cm) than area A (76.8 cm) over 1971-74, so that the means for weight expressed as a percentage of standard weight for length were 98.2% and 105.7% respectively (net weight). 1 These means are significantly different (p <0.001).

Weight and breeding success

In the first two parts of this section mean weights of adult females in three major breeding categories are compared, while in the third and fourth parts annual mean weights are compared and related to annual measures of breeding performance.

(i)

Overall mean weight and breeding success

Adult females that successfully reared young had the highest mean weight and the highest mean of weight as a percent of standard weight; adults not breeding had the lowest means, while those which lost young had means of intermediate value (Table 5). This suggests there is a relationship between stored fat reserves (weight) and breeding capacity.

1 This expression follows Bamford (1970) and is based on a pooled regression of log weight and log total length, the curve having the form W = 0.0163 l2.72 where W = weight (g) and l = total length (cm); a similar regression of log weight and log body length (computed from total length-tail length) was not used since the correlation coefficient was lower. For Westland possums Bamford's regression curve was W = 0.0125 l2.81; he found for each sex that weight as a percent of standard weight was highly correlated with the proportion of total body fat.

page 105
Fig 5. Overall annual and seasonal mean weights and mean weight/length indices for adult males and females area A (1966-75).

Fig 5. Overall annual and seasonal mean weights and mean weight/length indices for adult males and females area A (1966-75).

For clarity standard errors are omitted; for mean weights their range was 0.01–0.04 kg (males) and 0.01–0.03 kg (females); for mean weight/length indices 0.7–1.9% (males) and 0.7–1.1% (females). Only net weights of females are illustrated.

page 106
Table 5. Overall mean body-weights and mean body-weight/length indices for adult females in the three major breeding classes (net weights).
WITHOUT POUCH-YOUNG LOSING POUCH-YOUNG REARING POUCH-YOUNG
Mean N S.E. Mean N S.E. Mean N S.E.
WEIGHT (kg)
AREA A (1966-75): 2.13 298 0.02 2.19 519 0.00 2.34 663 0.01
AREA B (1971-74): 2.20 14 0.07 2.22 213 0.02 2.38 160 0.02
WEIGHT/LENGTH INDEX (%)*
AREA A (1966-75): 98.0 290 0.8 101.5 511 0.54 105.9 639 0.47
AREA B (1971-74): 94.3 11 3.2 95.9 213 0.82 101.7 160 0.61
(ii)

Seasonal mean weight and breeding success

Unlike adult males, adult females showed little overall seasonal differences in mean weight when data were pooled for all captures (Fig. 5). Thus in area A over 1966-75 means for net weight and net weight as a percent of standard weight were: 2.28 kg, 103.5% in autumn; 2.28 kg, 103.6% in winter; 2.26 kg, 102.7% in spring; and 2.27 kg, 103.4% in summer. In area B over 1971-74 there was more seasonal variation, females being heavier in autumn (2.37 kg, 101.2%) and summer (2.33 kg, 100.2%) and lighter in winter and spring (both 2.23 kg, 95.5%).

Seasonal variations in mean weight were more apparent when adult females were categorised according to their breeding success. In area A those rearing pouch-young had a relatively high summer mean which increased in the autumn to peak in winter, followed by a decline; females losing pouch-young were moderately heavy in summer, increased in mean weight in autumn (at the onset of the birth season) but then lost weight in winter, remaining relatively low in spring; those adult females without young had lowest means in summer which declined in autumn and winter (Fig. 6).

In area B over 1971-74 successful females gained weight in autumn as in area A, but then weights declined; those losing young were relatively heavier than the area A females in summer and autumn, but there was a marked decline by winter and a further decline in spring. As in area A, non-breeding adults weighed less than those with young, and weights again

* mean of body-weight expressed as the percentage of the standard body weight for length.

page 107
Fig. 6. Seasonal mean weight trends in adult females of different breeding histories.

Fig. 6. Seasonal mean weight trends in adult females of different breeding histories.

Comparable trends were evident when weight was expressed as a percentage of the standard weight for length.

Data for area A covers 1966-75; for area B 1971-74.

page 108 declined in winter (Fig. 6). Thus apart from the overall results suggesting a relationship between weight and breeding success (Table 5), more subtle seasonal relationships are evident which suggest that through various phases of the breeding cycle the female possum's breeding performance continues to be closely linked to her weight.
(iii)

Annual mean weight and breeding success

When mean weights based on the pooled capture data for each year were examined, a one-factor analysis of variance showed a significant heterogeneity in annual mean weights (net) of adult females in area A over 1966-71 (F ratio 32.49 with 9 and 2180 d.f., p <0.001). The annual weight means are shown in Table 6, together with annual mean lengths, again based on the pooled sample from all initial captures per month. Compared with the annual mean length range in males (76.0-79.8 cm), there was less annual variation in mean length of adult females (75.7-77.4 cm); thus the means of weight as percent standard weight tended to more closely parallel the raw weight means (Table 6), so both weight parameters are used as annual indices of mean fat reserves for females.

Annual weight means, when expressed as percent standard weight showed very similar fluctuations in both adult males and females (Fig. 5), suggesting both sexes were showing similar responses to annual changes in their environment. Changes in the age composition within the 'adult' age class could also influence annual weight and breeding trends, however; see, for example, breeding data in Table 2.

Annual mean weights for females are split into their seasonal components in Fig. 7, which shows there were different seasonal patterns of mean weight between years. These reflected some differences evident in Fig. 6. In 1971 in particular (also 1966, 1969) mean weights in area A increased in winter - in contrast to a marked decline in 1968 (Fig. 7). Again, increases occurred between summer and autumn in 1969 and 1971, but decreases occurred in 1968, 1973 and 1975. In area B a decline in mean weight occurred each winter over 1971-74, although autumn means were relatively high in most years (Fig. 7).

page 109
Fig. 7. Seasonal trends in mean weights of adult female possums: 1966-75.

Fig. 7. Seasonal trends in mean weights of adult female possums: 1966-75.

Weights are given as means for consecutive 3-month periods (± 1 standard error). Note general similarity of pattern between the 2 expressions of weight, and the variation between years, seasons and areas.

page 110
Table 6. Anuual variation in mean body-weight and length of adult female possums*

Table 6. Anuual variation in mean body-weight and length of adult female possums*

* based on combining all monthly estimates for each year

page 111
(iv)

Correlations between annual weight and breeding performance

The breeding performance of adult possums each year was significantly related to their mean annual body weight (Fig. 8). A comparison of the monthly mean weights of females and their breeding performance each year showed no correlations were evident over the summer months prior to breeding (Table 7). However, their weights over March and April, immediately prior to the autumn birth peak (Fig. 2), were significantly correlated with both the median date of birth and the proportion of females that bred (Table 7). The proportion of possums rearing pouch-young each year was most highly correlated with June to September mean weights, most young disappearing from the pouch over late August and September.

Table 7. Correlation of mean monthly body-weights (net) of adult females each year with breeding performance in those years, area A 1966-75.
MEDIAN DATE OF BIRTH PERCENT FEMALES BREEDING PERCENT POUCH-YOUNG SURVIVING
Significance1 r df Significance1 r df Significance1 r df
JAN NS .111 180 NS .029 180 NS .053 166
FEB NS .014 146 NS .059 146 NS .066 130
MAR *** .451 103 ** .301 103 NS .138 87
APR *** .248 193 ** .192 193 NS .096 169
MAY ** .238 165 * .169 165 * .199 139
JUN *** .439 181 *** .364 181 *** .363 165
JUL *** .362 205 *** .323 205 *** .243 196
AUG *** .513 163 *** .442 163 *** .366 136
SEP *** .432 203 *** .381 203 *** .308 190
OCT NS .103 165 NS .056 165 NS .141 149
NOV * .144 235 NS .057 235 ** .214 202
DEC NS .082 226 NS .042 226 NS .023 211
WHOLE YEAR *** .269 2187 *** .194 2187 *** .194 1962

1 NS = not significant;

*** = p<0.001

** = p<0.01;

* = p<0.05;

page 112
Fig. 8. Linear regressions and correlations between annual breeding performance and seasonal weights for adult female possums in area A (1966-75). Arcsin transformation for percent adults with young and percent rearing young.

Fig. 8. Linear regressions and correlations between annual breeding performance and seasonal weights for adult female possums in area A (1966-75). Arcsin transformation for percent adults with young and percent rearing young.

page 113

Even though most births occurred in May (Fig. 2), the correlations continued over June to September between mean monthly weights of females and both the proportion of them rearing pouch-young and the median date of birth (Table 7). This seems due to annual weight differences, already apparent in autumn, being even more marked and closely correlated in winter. The females' seasonal weight trends (Fig. 7) tend to support this: high weights in the autumns of 1966, 1969 and 1971 were even higher in the following winters; conversely, low weights in the autumns of 1968, 1970 and 1973 were lower by winter.

It was noted earlier that yearling females attempted to breed only in certain years (Table 4). In autumn yearlings averaged 1.23 ± 0.03 kg in area A but (like adults) the mean annual weights (± S.E.) fluctuated from 0.83 ± 0.13 kg (1974) to 1.55 ± 0.04 kg (1971). These yearlings' weights (Table 8) were correlated with adult female weights in autumn (Fig. 9). This suggests the factors affecting the condition of adult females during the birth season also influence the later development of young in their first year.

Table 8. Annual variation in mean autumn body-weights of yearling possums.
AREA A AREA A AREA B
YEAR MEAN ± SE(g) N YEAR MEAN ± SE(g) N MEAN ± SE(g) N
1966 1225±78 8 1971 1547±43 20 1293±97 7
1967 967±33 3 1972 1119±35 27 1282±72 14
1968 1083±87 6 1973 1058±123 6 1169±58 16
1969 1417±91 9 1974 825±132 5 1450±85 6
1970 1350±100 7 1975 979±161 6 -
Weights of pouch-young

The earliest a pouch-young was recorded free of the teat was at 53 days weighing 23 g (July 1972); of 443 weighings up to one year of age, only six were of young less than 90 days old. At 170 days, the assumed end of pouch-life, mean weights for 194 young up to 200 days old were estimated as:

page 114
(i)316 g using an age (X) and log weight (Y) regression Y = 0.0205X + 2.2741 (r = 0.8889);
(ii)328 g using an age (X) and cube root weight (Y) regression Y = 0.383X + 0.3809 (r = 0.8673).

Estimated weight at 170 days varied from year to year. Adopting the log weight regression (i) above, in area A, estimated weights were relatively high in 1969 (439 g), 1971 (369 g) and 1970 (352 g); were low in 1973 (264 g), 1974 (259 g) and 1972 (229 g); and nearer the overall area A mean (321 g) in 1975 (327 g), 1966 (310 g) and 1968 (297 g); only three young (mean 427 g) were weighed in 1967. In area B over 1971-73 the 170 day estimate was 291 g compared with 300 g for area A over the same period; annual means were 284 g (1971), 333 g (1972) and 256 g (1973). In area A over 1966-75 weight estimates at 170 days were correlated with the winter mean weight of 150 adults known to successfully rear pouch-young (Fig. 9). Thus years when female weights were higher were those when pouch-young weights were higher also.