The long-term study area includes the inshore and coastal waters along the central west coast of Florida, from Tampa Bay southward through Charlotte Harbor and Pine Island Sound, and in the Gulf of Mexico up to about 5-10 km offshore ( Figure 1 ). Most of the research effort, however, has been concentrated in Sarasota Bay and adjacent bays, sounds, and Gulf waters within about 1 km of the shore, due to initial findings of dolphin residency to these waters (Irvine and Wells, 1972). The shallow (< 4 m deep), sheltered bay and estuarine waters are separated from the Gulf of Mexico by a series of barrier islands, communicating with the Gulf through narrow, deeper (up to ~10 m deep) passes. The bays contain areas of shallow seagrass meadows and are fringed by mangroves, along with extensive manmade features such as bridges, piers, and seawalls. Natural or dredged channels 3-4 m deep run through seagrass meadows and sand or mud flats. A gently sloping, shallow sandy bottom extends offshore from the Gulf sides of the barrier islands.

The long-term resident Sarasota dolphin community inhabits a home range extending from southern Tampa Bay to Venice Inlet ( Figure 1 ; Wells, 2014). Sarasota and Manatee Counties, which encompass the Sarasota dolphin community’s home range, are heavily populated, with more than 900,000 people, and more than 46,000 registered vessels (as of 2023).

Longitudinal data on individually identifiable bottlenose dolphins were collected through observations, radio-tracking, sampling and measurements during catch-and-release operations, remote biopsy dart sampling, and stranding response.

More than 5,000 bottlenose dolphins have been individually identified by the SDRP along the central and southwest coast of Florida over the decades, but these are from multiple population units, and not all have been observed with sufficient frequency to be of value for analyses of life history, reproductive, and demographic parameters (Wells, 2014). For the analyses reported here, a subset of dolphins, belonging to an intensively and systematically studied, biologically meaningful population unit was selected, the Sarasota dolphin community (Wells, 2014). A dolphin community has been defined as “a regional society of animals sharing ranges and social associates, but exhibiting genetic exchange with other similar units” (Wells et al., 1999). Dolphins were considered to be residents of the Sarasota dolphin community if they were seen at least ten times and more than half of their sighting records occurred within the region bounded by Tampa Bay to the north and Venice Inlet to the south, and the barrier island chain to the west. In addition to facilitating the practical considerations of observation frequency, the geographical criterion also ensures that common evolutionary selection pressures are in force across the dolphin unit, shaping the parameters of interest.

The dataset included long-term resident Sarasota Bay dolphins that were seen with sufficient frequency that it was unlikely that important milestones in their lives, such as births and deaths, would be missed. Along the west coast of Florida, bottlenose dolphins live in long-term resident communities that form a mosaic, with slightly overlapping ranges (Wells et al., 1987; Urian et al., 2009). While regular systematic surveys that include waters adjacent to Sarasota Bay, along with radio-tracking, have demonstrated that Sarasota Bay residents often range into adjacent waters shared with other communities, for example along the Gulf of Mexico beaches or into southern Tampa Bay, the vast majority of Sarasota Bay resident sightings are inshore of the barrier island chain that defines Sarasota and associated bays (Irvine et al., 1981; Wells et al., 1987, 2013; Wells, 2014). Wells (2014) reported that, on average, 89% (± 12% SD) of the sightings of dolphins considered to be Sarasota Bay residents occurred within Sarasota Bay. Of those dolphins known to be at least 15 years old, 96% had been observed in the area over a span of at least 15 years, and some had been observed for as many as 45 years.

Dolphins were selected for analyses if they were observed during 1993-2019 and available data supported that they spent more time in the core portion of the Sarasota dolphin community range inside Sarasota and associated bays and sounds than outside of that region. The criteria included:

In total, 482 identifiable dolphins met the data-selection criteria. Of these, 69% were of known sex (179 f: 152 m: 151 unknown). The year of birth was known or estimated for 411 dolphins (85%), including individuals ranging in age from young-of-the-year calves up to 67 years. Fifteen dolphins were observed across all 27 years of the dataset; on average, individuals appeared in the dataset for 8.5 years. The dataset included 130 known mothers and 453 of their calves (some of these calves also became mothers). Reproductive histories continuing from the birth of the presumed first calf are available for 54 of the mothers in the 1993-2019 dataset, and include as many as 11 calves for any given mother (since these analyses were completed, a Sarasota resident was observed with her 12^th calf). For analyses involving full reproductive histories of mothers in the 1993-2019 dataset, calves born during 1973-2020 were included.

Dolphin reproduction is seasonal in Sarasota Bay, with most calves (81% of 353 assigned birthdates) born during May-July ( Figure 6A ). No births have been documented for January, and only one birth has been assigned to December. Spikes in reproductive hormones in blood, including testosterone for males and estradiol for females, occurred at the beginning of the peak of the calving season, consistent with a 12.5-month gestation period, although sampling did not occur during all months ( Figure 6B ). Seasonal changes in testis size consistent with the calving and hormone changes were also evident for adult males (≥10 years) ( Figure 6C ). Average testis length in winter was 72% of that in spring/summer (17.83 cm ± 4.13 cm sd, n = 6, vs. 24.66 ± 4.61 cm sd, n = 115; t-test, equal variance, 2-tailed, p = 0.0005). Four males were examined and sampled in both June and February. In all four cases, testosterone concentration, testis length, and testis diameter were greater in June. Declines in testosterone concentration and testis size appeared to occur in late autumn. Values were similar in both months for one male examined in June and October. A male examined in both June and November showed lower values for testosterone and testis diameter, but not testis length, in November.

Calving intervals varied greatly, depending on maternal age, birth order, and whether a calf survived long enough to separate from its mother. The shortest calving interval when the calf survived to separation was 1.9 years. The mean calving interval, regardless of calf survival to separation from mother, was 3.52 years (± 1.47 years sd, n=229), with a range of 1.02 to 10.05 years. Mean calving intervals when calves survived (n=105, 3.84 yr, CI 3.6-4.1) were significantly longer than when calves did not survive (n=124, 3.24 yrs, CI 3.0-3.5) (t-test for unequal variances, p=0.002). Mean calving intervals for successful cases preceding the birth of a male calf (mean = 3.67 ± 1.33 yrs sd, n = 28) were not significantly different from those preceding the birth of a female calf (mean = 4.18 ± 1.38 yrs sd, n = 35; t-test, 2 tails, p=0.14). Considering individual mothers, calving interval significantly increased with mother’s age ( Figure 7A ), whether a calf was successfully reared or not, and with birth order ( Figure 7B ) ( Table 2 ). No significant effect was found for calf sex. Testing each of those parameters individually to take advantage of larger sample sizes, significant effects were found for mother’s age and whether calves were successfully reared but not for birth order or calf sex ( Table 2 ).

The mean length of time that mothers invested in successfully raising a calf was 4.38 years (± 1.75 years sd, n = 143) for the maximum duration of association based on the COA and 4.23 years (± 1.72 years sd, n = 142) for abrupt separations. These estimates were not significantly different, however the variability surrounding both of these means was high. This variability reflects different calf-rearing approaches both among different mothers and for an individual mother. In the 159 cases for which post-separation associations with mothers could be examined over the remainder of the separation year, separation was complete for 88 calves (55.3%), indicating abrupt separation rather than gradual. These calves were from 78 mothers, of which 33 (42.3%) were only observed to engage in abrupt separations from their calves.

In 51% of the cases for which it was possible to determine, separation of a previous calf occurred prior to birth of the next calf. On average, prior separations (based on COAs) occurred 1.72 years (± 1.29 years sd, n=126) before the next birth. Post-birth separations of previous calves occurred on average 0.72 years (± 1.08 years sd, n=121) after the birth. Individual mothers did not consistently engage in one pattern or the other across calves.

There were no sex-related differences with regard to whether calves separated from their mothers before or after the birth of the next calf. Successful male calves remained with their mothers, on average, 4.42 years (± 1.56 years sd, n = 64), while successful females remained, on average, 4.43 years (± 2.01 years sd, n = 58). Differences by sex were not found in calf length or calf mass obtained within ±1 year of separation (both sexes combined: mean length = 211.5 cm ± 16.7 cm sd, n = 70; mean mass = 114.5 kg ± 28.5 kg sd, n = 66) ( Table 3 ).

Although there is much variability, the duration of association between mother and calf was most variable for younger mothers (the shortest and longest durations), with duration tending to increase with mother’s age, regardless of whether the separation was defined by COA values, or the timing of abrupt separation ( Figures 8A, B ). The pattern was similar for relative durations of association as a function of birth order, with the first births more variable ( Figures 8C, D ). Generally, this pattern appears to be manifested in calf sizes at separation, with calves being longer and heavier for younger and older mothers by the time they separate, and calves tending to be smaller at separation for mothers of intermediate age ( Figures 9A, B ). A similar pattern is evident for calf size at separation relative to birth order ( Figures 9C, D ).

Calves were observed after the death of their mothers, but before indications of separation, in nine cases. The durations of association with their mother for calves that survived beyond separation ranged from 1.2 to 8.9 years. The fact that a third of the orphaned calves survived on their own after a period of association with mother of less than 1.4 years provides an indication that nutritional support is of less importance to the prolonged mother-calf bond than other factors.

Lactation was noted for 81 females during health assessments. They were accompanied by calves up to nine years in age. Two of the lactating mothers had calves >6 yrs old, and 16% of females had calves >3 years old. In each case, the period of association between mother and calf continued beyond the date that lactation was recorded. Thus, considering that calves remained with their mothers for more than 4 years, on average, many were likely remaining with their mothers well beyond the time when they were receiving milk.

Over the 27 years considered for these analyses, the number of individuals present in any given year ranged from 108 in 1993 to 184 in 2018 ( Figure 10A ). On average, during any given year (1993-2019), 77.71% (± 12.37 sd) of groups were completely identified, and 87.86% (± 7.82 sd) of dolphins seen in the study area were identified. During the most recent decade, the percentage of identified dolphins increased to 95.28%. The catalog of resident dolphins was well-established at the beginning of the analysis period in 1993, with new individuals added at a relatively stable average annual rate of 14.4 dolphins/year (± 4.99 sd) thereafter, through reproduction, immigration, or fins that became distinctive ( Figure 10B ).

The age-sex distributions for 1993 and 2019 are shown in Figures 11A, B . The age-sex distribution of members of the community appears to be stable over time. A series of Kolmogorov-Smirnov tests found no significant differences between one year and the next, and no differences among the first year (1993), the middle year (2006), and the last year (2019) for which appropriate data were available.

In total, 268 births were documented during the 27-year study period. On average, 9.93 (± 4.39 sd, range 2-18/yr) calves were born each year. The mean annual birth rate was 0.071 births/dolphin (± 0.031 sd). It should be noted that, subsequent to compilation of the dataset used for these analyses, a record 22 calves were born to Sarasota residents in 2021.

The estimated number of mature females (6 years old and above) included in fecundity analyses in any given year ranged from 42 to 63. Mean annual fecundity was 0.182 (± 0.79 sd), ranging from 0.033 to 0.316. This value may be downward-biased by the low threshold for assigning maturity of 6 years, thereby likely including some females that were not yet mature, and by including females older than 48 years, beyond reproductive age. Removing females known to be older than 48 years of age yielded an estimated number of mature females (6-48 years old) in any given year ranging from 40 to 61, with a mean annual fecundity of 0.189 (± 0.81 sd), ranging from 0.036 to 0.321. It is possible that the dataset still included females of undetermined age older than 48 years.

Age-specific fecundity was examined for 241 calves born to known-age females ( Figure 12A ). In general, fecundity increases steadily from 7 years of age until females reach about 12 years of age, continues at a higher rate with no clear trend into the mid-20s, and then declines after about 25 years of age (Lacy et al., 2021).

Recruitment through reproduction was calculated based on 268 calves born to females considered to be resident during the 27-year analysis period. Of these, 205 (77%) were documented to have survived through at least their first year of life. The mean annual recruitment rate through reproduction was 0.050 (± 0.023 sd), ranging from annual rates of 0.006 to 0.097.

Non-reproductive recruitment was more difficult to define precisely, and included documented immigrants and possible immigrants, involving 10.8% of the residents. Over the 27-year analysis period, 11 of 482 resident dolphins (2.3%) were documented as having moved to Sarasota Bay after having been first documented outside Sarasota. Another 41 dolphins (8.5%) were first documented as residents when they were non-calves. Using the known immigrants to provide a lower bound, and the combined documented and possible immigrants as an upper bound, the minimum annual immigration rate was 0.003 (± 0.005 sd) dolphins per year, ranging up to a maximum annual rate of 0.013 (± 0.011 sd).

Of the 482 dolphins in the 27-year dataset, 129 were documented as mortalities, including 66 individuals greater than one year old and 63 younger calves. Of the 129, 78 were of known age and sex. The average age at death for the 43 females was 17.7 years (± 18.5 sd), with a median of 12 years. For the 35 males, the average age at death was 18.3 years (± 17.2 sd), with a median of 16 years. The oldest female in the confirmed death dataset was 58 years; the oldest male was 50 years. For both sexes, dolphins with older age estimates have been documented in Sarasota, but they disappeared (female = 67 years, male = 52 years).

The mean annual minimum mortality (i.e., known and presumed deaths) of all dolphins was 4.82 (± 2.32 sd) deaths/year, including 2.33 (± 1.82 sd) deaths/year for first-year calves, and 2.44 (± 1.63 sd) deaths/year for all others. The mean annual minimum mortality rate over all dolphins was 0.032 (± 0.014 sd), ranging from 0.006 to 0.051 in any given year. The mean annual minimum mortality rate likely underestimates the true mortality rate, as 67% of dolphins disappear without being recovered as carcasses or identified as having emigrated. Of the 482 dolphins in the dataset, 162 disappeared without further record over the 27 years of the analysis period. Combined with the known mortalities, the mean annual maximum mortality rate was 0.072 (± 0.025 sd), ranging from 0.034 to 0.132 in any given year.

Loss rates vary with dolphin age, with the highest rates occurring among the youngest and oldest age classes ( Figure 12B ). The rate of confirmed mortality from recovered carcasses for young-of-the-year calves was 0.067, but the actual value is likely 0.22, as disappearances of such young calves are also considered mortalities. Mortalities declined and remained low, at ≤ 1.5%, from the 6-10 year class through the 26-30 year class, then increased. The apparent decline in loss rate after 51-55 years is likely an artifact of small sample size for individuals reaching those estimated ages. Disappearance rates by age class track known mortalities well, suggesting that many of the disappearances with no other explanation could reasonably be considered to be mortalities.

For four of the 482 residents (0.8%, 3 males, 1 female), it was possible to document range shifts indicative of probable emigration. Three of these cases involved shifting to estuaries to the north, and one to the south. The mean annual minimum emigration rate, based solely on documented cases, was 0.001 (± 0.003 sd), ranging from 0.000 to 0.012 for any given year. Combining known emigrations with all disappearances of unknown cause as possible emigration events, the mean annual maximum emigration rate was 0.041 (± 0.022 sd), ranging from 0.000 to 0.088 in any given year.

Female reproductive success, measured as the proportion of calves seen post-separation, varies from none to all of their calves, with a mean of 0.45 (± 0.37 sd) for the 125 females for which information on calf presence post-separation was available. No surviving calves were identified for 28% of the 125 mothers, while all calves are known to have survived for 22% of the mothers. Two Sarasota females have successfully reared at least six calves each.

Overall, of 426 calves with documented birth year and known or presumed fates, including calves born prior to 1993, 78.2% survived their first year, 61.1% survived their first three years, and 48.4% were identified post-separation. Maternal experience appeared to play a role in reproductive success, as exemplified by analyses of the fates of 224 calves of known birth order. Only 62.3% of first-born calves survived their first year as compared to 79.5% of subsequent calves (Fisher Exact Test, chisq p=0.0168). Calves of multiparous females also showed higher, but not significant, survival at three years (60.5% multiparous, 54.7% primiparous, Fisher Exact Test chisq p=0.5205), and was near parity at post-separation (54.8% and 50%, Fisher Exact Test chisq p=0.8753).

Male reproductive success could be assessed for 52 cases where males were identified as the only possible sire for calves born to resident Sarasota females. The maximum number of calves sired was seven, documented for two males. Most of the calves were sired by males in their twenties to mid-thirties, with no males older than 43 years identified as sires ( Figure 12C ).

Some results from the Sarasota Bay study are not available from other dolphin longitudinal studies, e.g., lifetime reproductive success and longevity. However, other long-term studies have provided observational data to estimate, or place bounds on, age-specific demographic parameters that require long time-series as well as ages of known individuals. Age at first birth, or the related parameter, age at sexual maturation (ASM), is a parameter often estimated. For 11 female Indo-Pacific bottlenose dolphins observed in western Australia during 16 years, first birth for 4 females occurred at 12-15 years of age, while 7 females of the same age had not yet given birth (Mann et al., 2000). Similarly, in a 16-year study of common bottlenose dolphins in Doubtful Sound, NZ, 3 females of known age gave birth at 10.8-11.8 years of age, while 4 females over the age of 12 years had not yet calved (Henderson et al., 2014). In a 20-yr study of common bottlenose dolphins in Scotland, of 16 females known since birth, 11 gave birth at ages 8-9, while the youngest and oldest primiparous females were 6 years (n=2) and 13 years (n=1), respectively; with the assumption that no earlier births by the older female were missed, the mean was 8.2 years (Robinson et al., 2017). In a 16-year study in eastern Scotland, of 13 females sighted every year since birth, age at first birth ranged from 6-14 years (median=9) (Cheney et al., 2019). Indo-Pacific bottlenose dolphins under human care were determined to be mature on the basis of a detected ovulation or hormone levels as early as 6 years of age at one facility (Zhang et al., 2021; Brook, 1997), while first conception occurred at 11-13 years in a 10-year study at different facility, with age at maturation unknown (Cheal and Gales, 1991).

Although a different genus, for 22 female Amazon River dolphins observed during a 24-year study, first birth occurred at 7.2-12.4 years (mean=9.7), with 3 of the 76 mature females never having been seen with a calf (Martin and Da Silva, 2018). These results are similar to those from Sarasota Bay (for 54 resident females observed over 38 years, first birth ranged from 6 to 14 years, mean=9.6, with four older females never having been seen with a calf), resulting in Martin and Da Silva (2018) suggesting that some life-history traits are conserved; this similarity maybe be correlated with body size (Calder, 1996; Peters, 1986). The two longer-term studies (current study and Martin and Da Silva, 2018) were able to put upper bounds on age at first birth, while the shorter-term studies described above had not yet identified that bound, illustrating the length of study time needed to estimate age at first birth in late-maturing, long-lived species. They also illustrate the late age at which dolphins of this size become mature, on average, and the variation in age at first birth, with a range for Inia of 5 years and, to date, 8 years for Tursiops. These studies also illustrate the older ages at which some females have not yet given birth.

More data, and increased sample sizes, are available for parameters requiring a shorter time frame to observe and ages known only for calves. Two related parameters often estimated are calving interval (CI) (or inter-birth interval), and age at weaning. The CI is reported differently among studies, e.g., including or excluding intervals when a calf did not survive to weaning, limiting some direct comparisons and introducing different biases (Arso Civil et al., 2017; Cheney et al., 2019). Nonetheless, patterns emerge among studies. The CI ranges from as low as one year, when a calf was lost shortly after birth to as high as 10 years (current study, Mann et al., 2000; Henderson et al., 2014; Robinson et al., 2017) and is influenced by calf survival, mother’s age, birth sequence, timing of birth, group size, prior successful births, and other factors (current study, Henderson et al., 2014; Blasi et al., 2020; Robinson et al., 2017; Bezamat et al., 2020). Direct observations of a CI of 2 years for females successfully weaning calves are rare but have been reported (current study, Arso Civil et al., 2017; Bezamat et al., 2020; Cheney et al., 2019; Steiner and Bossley, 2008), as were CIs greater than seven years (current study, Mann et al., 2000; Arso Civil et al., 2017; Henderson et al., 2014; Robinson et al., 2017), all of the latter being recorded in field studies of greater than 17 years. A shorter range of observed CI was reported for shorter-term studies (Indo-Pacific bottlenose dolphins, range 1-6 years over an 8-yr period, Kogi et al., 2004; common bottlenose dolphins, range 1-2 years over two disjointed periods for a total of six years, Bezamat et al., 2020; common bottlenose dolphins, range 2-7 years over an 8-yr period, Baker et al., 2018), which may reflect actual differences in range of CI for those populations or be biased downward given the length of the studies (Barlow and Clapham, 1997; Arso Civil et al., 2017; Cheney et al., 2019).

Excluding the shorter-term, potentially negatively biased, studies noted above, the mean CI was similar among studies of bottlenose dolphins, ranging from 3.3 to 4.7 yrs (current study, Mann et al., 2000; Arso Civil et al., 2017; Henderson et al., 2014; Blasi et al., 2020; Robinson et al., 2017; Fruet et al., 2015). The lowest mean CI was 1.7 years, estimated from all birth intervals including lost calves, for a sample of five reproductive females observed during a 16-yr study, while CI for surviving calves of nine females was 3.8 years (Steiner and Bossley, 2008). When provided, the mean CI differed when estimated from all calving intervals or from intervals only when a calf survived, although the direction of difference varied among studies. The studies recognized that the mean CI will be positively biased if non-surviving calves were missed, which will be affected by the intersection of survey frequency and calving seasonality. The average age at weaning, when estimated, was the same as, or close to, CI for all of the studies. It should be recognized, however, that age at nutritional weaning and age at separation are likely to not be the same, as the period of association between Sarasota Bay mothers and calves continued post-lactation, a difference often not distinguished, and possibly not known, in most published studies.

The CI has been estimated for other small-cetacean species from long-term studies, as well. In a 24-yr study of Inia, the CI ranged from 2 to 7.7 years with a mean of 4.6 when including all calves sighted (Martin and Da Silva, 2018). Duration of nursing ranged from 1.5 to 5.8 years (mean=2.8 years). In an 11-year study of Stenella frontalis, the CI ranged from 1 to 5 years with a mean of 3.6 (all calves included) or 4.3 years (including only when a calf successfully weaned) (Herzing, 1997). In a 13-year study of beluga whales (Delphinapterus leucas), CI ranged from 2 to 13 years (McGuire et al., 2020). Zeng et al. (2021) estimated a range of 3.5 to 5.8 years for Sousa chinensis in a study lasting six years, suggesting the possibility of a downward bias in CI due to the relative length of the study.

Many studies report data pertaining to pregnancies or births, including crude or mean annual birth rate (the number of calves relative to the total number of individuals less the number of births) (e.g., Steiner and Bossley, 2008; Baker et al., 2018), fecundity, annual fertility rate (e.g., Pleslić et al., 2019; Cheney et al., 2019), and annualized birth rate when fecundity is corrected for length of gestation (e.g., Martin and Da Silva, 2018). Fecundity has variously been defined as the proportion of calves surviving to age 1 year relative to the number of mature females (also called annual recruitment rate) (e.g., Herzing, 1997; Zeng et al., 2021), annual proportion of females giving birth (e.g., Arso Civil et al., 2017; Baker et al., 2018), and the number of female offspring relative to the number of mature females (e.g., Fruet et al., 2015; Robinson et al., 2017). Direct comparison across studies is not possible because the parameters estimated and reported are non-standardized. However, given an unbiased estimate of CI (see Arso Civil et al., 2017), fecundity is the reciprocal of the CI and may be more readily estimated as such. For the various long-term studies, the reciprocal of the mean CI results in estimates of fecundity of 0.21 to 0.30, which is somewhat higher than the estimates of fecundity for the Sarasota Bay community (range 0.12-0.19, depending on estimation method). However, among studies, when both fecundity and CI are reported, the reciprocal of CI is often higher than the estimates of fecundity (current study, Kogi et al., 2004; Herzing, 1997), and can be incomparable (e.g., Bezamat et al., 2020; Fruet et al., 2015; Martin and Da Silva, 2018), which likely reflects the various ways the parameters are estimated or is due to undetected fetal or neonatal loss. Martin and Da Silva (2018) report a large difference in CI and pregnancy rate in Inia; however, if every detected pregnancy (via ultrasound) resulted in a detected calf, then the two parameters would have been comparable. For useful comparisons among populations, estimation methods for CI and reproductive parameters would need to be better standardized. That standardization would be especially helpful for parameters that can be estimated from shorter-term studies and that are used in population models, such as the use of age-specific reproductive rates in population viability analysis (PVA) (e.g., Lacy et al., 2021) or for correcting for data-driven biases (Arso Civil et al., 2017).

Reproductive success and output vary within communities, albeit with some commonalities among studies. Variation occurs in the number of calves per mature female and the survival of those calves, with some mature females not seen to successfully raise a calf to separation (full independence) while other mature females raise all of their calves to separation (e.g., current study, Henderson et al., 2014; Fruet et al., 2015). Further, primiparous females have been found to be less successful at raising calves to separation and multiparous females more successful, in general, although some primiparous females are more successful than others (current study, Reddy et al., 2001; Wells, 2000; Schwacke et al., 2002; Baker et al., 2018; Henderson et al., 2014; Blasi et al., 2020; Robinson et al., 2017; Cheney et al., 2019; Mann et al., 2000; Arso Civil et al., 2017). Most longitudinal studies report females having three or fewer calves, while the longer-term observations have resulted in a higher number (Robinson et al., 2017; Arso Civil et al., 2017), including in Sarasota Bay, where females have been observed through as many as 12 calves. While data on age-specific fecundity are limited due to the duration of longitudinal studies and knowledge of individual ages, reproductive output has been reported to increase initially, stabilize, then decline in older females (current study, Lacy et al., 2021; Robinson et al., 2017, and Perrin et al., 1977 from carcass studies). Adjustment of results from shorter-term studies, e.g., that extrapolate from primarily younger females, may not capture the full variation in the population nor age-specific differences in calving success or survival of mature females giving birth, nor account for reproductive senescence in older females. A population viability analysis (PVA) for Sarasota Bay dolphins found that population growth was most influenced by reproduction in peak breeding-age females (ages 12-25 years), along with calf and juvenile survival (Lacy et al., 2021), illustrating the importance of robust estimates of age-specific reproduction.

Examination of carcasses can provide data on some reproductive parameters, although such results are not necessarily comparable to data from observational or catch-and-release studies. For large and presumably unbiased samples, pregnancy rates can be estimated on the basis of the relative number of fetuses (e.g., Perrin and Donovan, 1984; Murphy et al., 2009). Otherwise, counts of ovarian scars from (presumed) pregnancies (corpora lutea or albicantia) have been used to estimate the number of possible births (see Perrin and Donovan, 1984; Murphy et al., 2009), although the interpretation may be complicated with species differences (Dabin et al., 2008). A valuable opportunity to compare counts of ovarian corpora to observed births in three free-ranging female Indo-Pacific bottlenose dolphins with known life histories resulted in an equal number of corpora and births (Kemper et al., 2019), and for a female Indo-Pacific bottlenose dolphin under human care and undergoing frequent ultrasound examinations the number of corpora albicantia equaled the number of calves born (n=3) and not the number of ovulations (n=18) (Brook et al., 2002). There was no opportunity to evaluate corpora persistence following fetal loss. However, caution is needed in translating pregnancy rates from carcasses to successful reproduction. For example, Martin and Da Silva (2018) found that the pregnancy rate, as detected using ultrasound, was twice the annual birth rate (0.22) based on the sighting of calves. While some calves may have been missed during visual surveys, the difference between the two parameters suggests significant fetal loss. In Sarasota Bay, 83% of fetuses observed in ultrasound exams were subsequently observed as living calves, suggesting a fetal loss of up to 17% (Wells et al., 2014). Further, fetal loss, as well as loss of the first calf, is amplified by high contaminant burdens in mothers, including exposure to chemicals from oil spills (e.g., Wells et al., 2005; Schwacke et al., 2002; Reddy et al., 2001; Lane et al., 2015) and likely would vary among geographic areas. For example, in Barataria Bay, Louisiana, where bottlenose dolphins were exposed to the Deepwater Horizon oil spill, only 20% of fetuses were subsequently observed as calves (Lane et al., 2015). As a result, calving rates may be overestimated, and age at first birth may be underestimated in studies from carcasses, as lost fetuses from early pregnancies would delay successful births, thus decreasing calving rates and increasing the actual age at first birth. While data-based biases occur for both observational or catch-and-release studies and for studies using carcasses, the biases are different, and studies from carcasses may suggest higher population growth rates or potentials than are likely.

While age at first birth is more important for estimating population growth, for populations with short-term or no longitudinal studies, which is most populations of cetaceans, the ability to estimate ASM from carcasses may be useful. ASM is more conservative and less subject to the range of environmental influences than age at first calving (Millar and Zammuto, 1983; Stearns, 1983; Ferguson and Higdon, 2013; Peters, 1986; Calder, 1996; 1996, Sibly and Brown, 2007). In a representative sample, this estimate may be more accurate than the estimate from observational studies, which would be age at first birth minus the gestation time, given the evidence of fetal and neonatal losses in younger females. However, given actual or potential biases in carcass samples, results from longitudinal studies may provide guidance for when a sample is adequate for estimating a robust ASM from carcass samples. In particular, one result from estimates of age at first birth from observational studies is that, within a population, there is a protracted range of ages over which first birth (and resulting ASM estimated from age at first birth) occurs. This results in a number of indeterminate age classes, which are age classes for which maturity cannot be assigned by age alone. Samples not including the full range of indeterminate ages are likely biased, with many sample age distributions tending toward younger animals or smaller size (e.g., Barlow and Hohn, 1984; Moore and Read, 2008; Venuto et al., 2020; Berner and Blanckenhorn, 2007), which would result in a downward bias. In these cases, the sample is simply inadequate to obtain a robust estimate of ASM.

As has been reported for several species of small cetaceans, apparent reproductive senescence was documented for 10% of all known-age mothers in Sarasota Bay, with the onset ranging from 25 to 48 years of age. In short-finned pilot whales (Globicephala macrorhynchus), females live many years beyond the age of last reproduction (Kasuya and Marsh, 1984). In resident killer whales (Orcinus orca), reproductive senescence can last more than 50 years, starting at 40 years of age (Foster et al., 2012). Post-reproductive females can serve as repositories of important information gleaned from decades of experience, including such things as knowledge of resources and how to acquire them, predators and how to defend against them, and social knowledge that can help in navigating social interactions (grandmother hypothesis, Johnstone and Cant, 2010; Croft et al., 2017). Further, the existence of post-reproductive females may increase cooperation behavior (Ross et al., 2015) and stabilize population dynamics (Mitteldorf and Goodnight, 2012). It has been demonstrated that post-reproductive female resident killer whales increase the fitness of their kin through transfer of ecological knowledge (Ward et al., 2009; Brent et al., 2015). In resident killer whales, direct application of this information to the benefit of long-term, related swimming associates is possible. The evolutionary advantages derived from the continued existence of Sarasota Bay females in their fluid society well beyond their reproductive years remain to be determined.

Adult male Sarasota bottlenose dolphins can be easily distinguished from immature males based on testosterone concentrations in serum and blubber, with adult males having 300 times greater mean serum concentration than immatures, and 100 times greater in blubber (Sherman et al., 2021). Testis volume was highly correlated with blubber testosterone, with mature males having 60 times greater volume than immatures (Sherman et al., 2021). The short period of time over which maturation occurs means that few males would be expected to have been indeterminate with respect to maturation.

Estimates of age and length at maturation in male dolphins from long-term observations are rare. Over 10 years of observations, three male Indo-Pacific bottlenose dolphins under human care were estimated to be 11 to 12 years old when they first sired calves (Cheal and Gales, 1991). In contrast, using the first presence of sperm in semen to define sexual maturity, five Indo-Pacific bottlenose dolphins monitored for 3 to 5 years matured at ages 6.6 to 7.3 years (Yuen, 2007). Estimates of age at first siring and age at first sperm production for a population are quite different parameters given the delay in siring for males, and, as noted earlier for females, neither is likely to be directly comparable to the ASM estimated from carcasses. Although a mean age at first siring is not available, age at maturation for Sarasota Bay males would be less than age at siring (i.e., social maturation), given that the youngest sires are 10 years of age at conception and most siring occurs when males are in the mid-teens and older.

Sarasota males matured later than females, both for age at first reproduction and ASM. Length and mass at maturation in males were not different from age at first calving and first birth in females, although these values are not equivalent as females would have an additional year of growth, and thus, it would be expected that they were smaller one year prior. Male-biased sexual dimorphism in asymptotic size occurs in most delphinids (Caspar and Begall, 2022; Ralls and Mesnick, 2009; Tolley et al., 1995; Moller et al., 2007; Möller, 2012) and does occur in the Sarasota Bay dolphins (Read et al., 1993). Most of the data supporting dimorphism in age or length at maturation are from carcass studies (e.g., Murphy et al., 2009; Betty et al., 2022; Ngqulana et al., 2017) rather than studies of free-ranging individuals; in carcass studies, direct comparisons of dimorphism would be comparable provided the sample size is sufficient.

Law (1956) proposed that sexual maturation in male delphinids is attained at 93% of total length. Sergeant et al. (1973) proposed that it occurs at 90% of total length in bottlenose dolphins. Cheal and Gales (1991) documented maturation to occur in male Indo-Pacific dolphins under human care coincident with the secondary growth spurt. The value from Sarasota Bay dolphins is similar to other published values, with predicted LSM occurring at 89% of asymptotic length. Body size and age at maturation are linked (Stamps et al., 1998; Sibly and Brown, 2007) so using a robust growth curve may be helpful to evaluate whether an ASM estimate is reasonable.

After sexual maturation, there is significant variability in serum testosterone levels and testis dimensions relative to age and body size in Sarasota Bay males. Variability in these parameters has also been documented in dolphins under human care (Yuen, 2007; Yuen et al., 2009; Brook, 1997; Schroeder and Keller, 1989). In three of four “elderly” Indo-Pacific bottlenose dolphins under human care ranging from 47 to 50 years of age, Koga et al. (2019) found testosterone levels and sperm concentrations to still be relatively high and suggested that even older dolphins retained their ability to breed. The results from mature males in Sarasota Bay were similar; however, there is an apparent decline in testosterone after about age 30, as well as a decline in siring frequency a few years earlier. The oldest documented actively breeding male in Sarasota was 43 years old, 9 years younger than the oldest documented male, suggesting that successful breeding does not continue through a male’s entire adult life. Whether the decline in siring is due to social competition or insufficient production of sperm or semen is unknown. The potential role of accumulated environmental contaminants in the decline is also unknown, as concentrations of some of these are known to exceed thresholds for health and reproductive impacts in some older Sarasota males (Wells et al., 2005).

Seasonality in reproductive parameters of male dolphins is common among sampled populations (Plön and Bernard, 2007). Studies on Tursiops spp. under human care have found elevations in testosterone levels, sperm counts or density, and/or other metrics over periods as brief as a couple of months during the year (Schroeder and Keller, 1989; Mingramm et al., 2019; Katsumata et al., 2017; Koga et al., 2019; Yuen, 2007; Wells, 1984; Funasaka et al., 2011). Schroeder and Keller (1989) have shown that peak testosterone levels preceded peak sperm density in bottlenose dolphins under human care. In Indo-Pacific bottlenose dolphins under human care, Brook (1997) and Yuen (2007) found a correlation between testosterone levels and testis volume that was cyclical, coinciding with ovulation in females, although Yuen (2007) found no annual pattern in sperm density. Seasonal changes in testosterone levels in blubber samples have been demonstrated from free-ranging bottlenose and short-beaked common dolphins (Sherman et al., 2021; Kellar et al., 2009; Boggs et al., 2019). While only limited sampling in Sarasota Bay occurred outside of the breeding season, the results are consistent with a seasonal increase in testosterone, coincident with breeding season (Sherman et al., 2021). The use of ultrasound also allowed detection of a decrease in testis size with lower testosterone levels outside of the breeding season, as was reported for dolphins under human care.

Determining maturation and seasonality of reproduction in small cetaceans has predominantly been based on carcass studies. Studies with an adequate sample size, especially in indeterminate age classes for sexual maturation, have confirmed the dimorphism in average ASM and LSM between males and females. Two differences are relevant with regard to defining ASM and LSM in males and females, however. First, while a female is considered to be clearly mature when at least one corpus of ovulation is present on at least one ovary, the precise timing of maturity in males is ill-defined with no single criterion widely applied (Perrin and Henderson, 1984, Plön and Bernard, 2007; Yuen, 2007). Applying models that estimate or predict ASM or LSM using the proportion mature in each age class based on criteria generally known to indicate maturity, such as testosterone levels, testis size, or sperm presence, should still provide reasonable estimates for males if the sample is sufficient. Second, for longitudinal studies such as in Sarasota Bay, ASM or LSM will not be known for females based on first ovulation and will need to be estimated from age at first birth, which, as discussed above, is likely to overestimate ASM. In contrast, estimates for males are likely to be comparable to those from carcass studies. Thus, dimorphism in average ASM and LSM may be underestimated and not directly comparable to estimates determined for males and females from carcasses.

Seasonal changes in testis dimensions, hormone levels, or histological features have been shown for a number of small-cetacean species from studies using carcasses (Hohn et al., 1985; Read, 1990; Slooten, 1991; Van Waerebeek and Read, 1994; Hohn et al., 1996; Fisher and Harrison, 1970; Collet and Saint Girons, 1984; Neimanis et al., 2000; Desportes et al., 1994). In addition, studies from carcasses of long-finned (Globicephala melas) and short-finned pilot whales, allowing for morphometric and histological examination as well as measurement of testosterone levels, showed that testosterone levels were not correlated with testis mass (Kita et al., 1999) or body length or age after sexual maturation (Desportes et al., 1994). Results from a large sample of pantropical spotted dolphin carcasses showed that maximum testis size peaked prior to the predicted mating season, while spermatozoa levels were elevated during the breeding season (Hohn et al., 1985). These estimates are more straight-forward than estimating reproductive parameters for females, as testis size and spermatogenesis should provide a non-biased snapshot for males if the sample size and seasonal distribution of samples are adequate, and given the constraint that maturation in males is not uniformly defined. For those parameters measured in longitudinal studies and from carcasses, however, the results on seasonality should be generally comparable.

Sarasota males were determined through paternity testing to be sires of up to seven calves each within the Sarasota dolphin community (and they could have sired more beyond Sarasota; Duffield and Wells, 2023). Members of strongly bonded male pairs sired 75% of calves, and individual males within pairs were closely matched for age, length, mass, and association history with the females (Owen et al., 2002; Duffield and Wells, 2023). Duffield and Wells (2023) examined characteristics of males seen in association with females during periods of presumed female receptivity to identify features associated with breeding success. Breeding males were characterized by age, in that 77% of sires were older than the mothers, on average exceeding mother’s age by more than five years, in spite of the fact that females can still produce calves at 48 years of age, and very few males live that long (Lacy et al., 2021; Wells, 2000). Successful sires were significantly older and heavier than non-sire associates during the period of female receptivity (Duffield and Wells, 2023). Taken together, these observations suggest that some form of male-male competition and male dominance may be in play, within a polygynandric system that involves mate-guarding.

The demographic structure and dynamics of the Sarasota community can serve as a model for other cetacean communities. The number of Sarasota dolphin community members identified in any given year has been variable, ranging from 108 in 1993 up to 184 in 2018 ( Figure 10A ). This increase occurred in large part after a 1995 banning of fishing nets in inshore Florida waters that captured many of the dolphins’ prey fish species and occasionally killed dolphins. Declines from one year to the next often coincided with or followed major environmental perturbations, such as severe harmful algal blooms (Karenia brevis red tides) that depleted prey and/or increased predation pressure or increased adverse human interactions (Gannon et al., 2009; Berens McCabe et al., 2021; Wells et al., 2019). In recent years, the increase in numbers of identifiable Sarasota dolphin community members appears to have slowed, suggesting that the community may be reaching its carrying capacity (Lacy et al., 2021). The current community size of up to about 200 individuals including both identifiable and unidentifiable dolphins is comparable to sizes identified for a number of other bay, sound, and estuary regions along the Gulf coast of Florida. For example, the estimated 524 dolphins of Tampa Bay were spread among four different communities (Urian et al., 2009). Multiple communities likely exist in the Charlotte Harbor and Pine Island Sound region, where about 826 dolphins were estimated to occur (Bassos-Hull et al., 2013). A resident community of 78-152 dolphins was identified in St. Joseph Bay (Balmer et al., 2008).

Some vital rates, such as immigration and emigration, are difficult to measure in the absence of long-term study and data collection beyond the geographic range of the focal community, so data for comparison to Sarasota are lacking. The low estimated mean rates of immigration (0.003-0.013), emigration (0.001-0.041), and occasional movements of breeding males among communities are sufficient to maintain genetic exchange with neighboring communities (Duffield and Wells, 2023). Data on parameters such as estimated birth rates and mortalities are generated from within a community, providing more opportunities for comparisons. The Sarasota mean annual birth rate of 0.071 falls within the range of previously reported bottlenose dolphin values of 0.012 to 0.156 (Leatherwood and Reeves, 1982; Perrin and Reilly, 1984). For Sarasota dolphins across all age classes, mean annual mortality rates were estimated as 0.032-0.072, recognizing that the upper limit is biased upward due to including losses from all causes, including undocumented emigration events and changes in identifying characteristics. Comparable or higher loss rates have been reported from other sites. Off Charleston, South Carolina, the estimated annual loss rate is 0.049, without being able to distinguish between mortality and permanent emigration (Speakman et al., 2010). Ludwig et al. (2021) estimated an annual loss rate of 0.06 for the Shannon River estuary in Ireland. Stolen and Barlow (2003) estimated an overall annual mortality rate of 0.098, based on stranding data. In the southwestern Gulf of Mexico, estimated annual loss rate was about 0.12 (Bolaños-Jiménez et al., 2021). Around Reunion Island, the apparent annual loss rate was 0.17 (Estrade and Dulau, 2020). For perspective, a loss rate of 0.132 was estimated for the Barataria Bay, Louisiana, population of bottlenose dolphins impacted by the Deepwater Horizon oil spill (Lane et al., 2015).

Calf losses varied with maternal experience, and from site to site, with young-of-the-year typically suffering the highest losses. Overall, of 426 Sarasota calves with documented birth year and known or presumed fates, 78.2% survived their first year, 61.1% survived their first three years, and 48.4% were identified post-separation. Only 62.3% of first-born Sarasota calves survived their first year, as compared to 79.5% of subsequent calves. Along the eastern Adriatic coast, 87.5% of calves survived their first year (Pleslić et al., 2021). In Doubtful Sound, New Zealand, 67% of calves survived their first year, and 40% survived through their first three years (Henderson et al., 2014). In Bay of Islands, New Zealand, first-year mortality was 0.34-0.52 (Tezanos-Pinto et al., 2015). First-year Lahille’s bottlenose dolphin calves (T. t. gephyreus) were lost at a rate of 0.16 (Fruet et al., 2015). Environmental contaminants have been suggested as a potential contributor to elevated losses of first-born calves (e.g., Schwacke et al., 2002; Wells et al., 2005).

Population growth rates have been estimated for a few bottlenose dolphin populations. Lacy et al. (2021) estimated a long-term (27 years) mean annual population growth rate of 0.021 (sd = 0.031%) for the Sarasota dolphins, noting that population growth was most sensitive to uncertainty and annual variation in reproduction of peak breeding-age females and in calf and juvenile mortality, with little variation in adult survival over time. Stolen and Barlow (2003) estimated a rate between 0.00 and 0.046 for Tamanend’s bottlenose dolphin (T. erebennus) from a life table based on stranding data in the Indian River Lagoon of Florida. A growth rate of -0.007 was reported for the 145-dolphin population in the Shannon River Estuary of Ireland (Blázquez et al., 2020). Given the slow rates of population growth for bottlenose dolphin populations, it is estimated that the population of dolphins in Barataria Bay, Louisiana, will require 35 years to recover from the Deepwater Horizon oil spill (Schwacke et al., 2022). It was estimated that a resident estuarine population of Lahille’s bottlenose dolphin in southern Brazil would grow at a rate of about 0.03 annually in the absence of bycatch (Fruet et al., 2021). Most measured growth rates are less than the predicted growth rate of 4% commonly used for delphinid population growth (Reilly and Barlow, 1986).

In the late 1960s-early 1970s, bottlenose dolphins, typically young females, were collected from many parts of the southeast U.S., including Sarasota waters, for public display, research, and military applications (Leatherwood and Reeves, 1982). There was little appreciation at the time that these populations of bottlenose dolphins were relatively discrete, small and resident to the coastal bays and estuaries. As such, removals that targeted specific bays and specific age cohorts had the potential to create a long-lasting gap in the age distributions. Gaps were apparent in the age distribution (see Figure 11 for the year 1993) that coincided with the collections, projected through time. It might be expected that this gap, as it projected through the prime female breeding ages, could have had effects on calf production. No such effect was observed, however, likely due to the confounding effects of high annual variability in calf production, a changing carrying capacity after an inshore gillnet fishing ban was enacted in Florida waters in 1995, and developing climate change.