Female reproductive output in a Mediterranean bottlenose dolphin Tursiops truncatus population

The reproductive output of wild bottlenose dolphins Tursiops truncatus has extensively been investigated; nevertheless, it remains widely unexplored in the Mediterranean Sea. We investigated the reproductive history of female bottlenose dolphins in the Aeolian Archipelago (Italy). Thirteen years of photo-identification data were used to track the reproductive pattern of 7 females and their 16 calves. The birth period was accurately estimated within 1 mo for 6 calves (37.5%), peaking in June/July. The mean number of calves for females was 2.3 ± 1.1 (SD), ranging from 1 to 4. The mean interbirth interval was 3.5 ± 1.6 yr, ranging from 2 to 7 yr. The mean weaning age was 3.2 ± 0.6 yr, although 1 calf may have been weaned as late as its fifth birthday. Mother−calf postweaning associations were tracked for 9 calves and ranged from 4 to 10 yr of residence in the maternal group of the calf. The age of first pregnancy was 7.5 yr. The number of calves was influenced by the mean maternal group size, with females that were sighted in larger groups having more calves. However, the number of calves was not influenced by the strength of the associations of mothers with other females. Calf mortality by age 1 was high (12.5%). Two juveniles also died postweaning within 6 yr of their birth. Finally, 2 females may have died during the study, as they were not re-sighted at the end of the study. Our results highlight the need to implement urgent conservation actions to protect this bottlenose dolphin population.


INTRODUCTION
The reproductive output of wild bottlenose dolphins Tursiops truncatus has been extensively explored worldwide (Fruet et al. 2015, Rossi et al. 2017, Baker et al. 2018, especially for resident populations inhabiting estuary areas of sub-tropical regions (Shane et al. 1986, Steiner & Bossley 2008, Bezamat et al. 2020; however, it remains largely undocumented in the Mediterranean Sea, where bottlenose dolphins have been studied only in relatively small areas (Gol'din & Gladilina 2015, Rossi et al. 2017. Bottlenose dolphins live in fission− fusion societies characterized by ex tensive maternal investment, with prolonged mother− infant relationships and juveniles occasionally associating with their mothers post-weaning (Oftedal 1997, Mann & Smuts 1998. Common reproductive trends have been identified among different bottlenose dolphin populations (Henderson et al. 2014, Brough et al. 2016), but some reproductive parameters, such as seasonality of births (Urian et al. 1996, Thayer et al. 2003, Karniski et al. 2018, weaning period (Tavolga 1966, Cockcroft & Ross 1990, Mann & Smuts 1998, and age of first reproduction (which may vary between 7 and 12 yr for bottlenose dolphins, depending on region), may vary greatly among populations (Wells et al. 1987, Schroeder 1990, Kasuya et al. 1997) as a result of ecological and social factors, e.g. group size, protection from predators/conspecifics, and mate choice , Krzyszczyk et al. 2017).
Photo-identification (photo-ID) is a technique which uses scarring patterns and unique nicks on the dorsal fins of individuals to investigate the ecology and behaviour of dolphins (Würsig & Jefferson 1990). Several studies have applied photo-ID methods to investigate demographic and reproductive factors in wild bottlenose dolphin populations (Grellier et al. 2003, Robinson et al. 2017, Bezamat et al. 2020. Hammond et al. (1990) used the individual reproductive history of repeatedly sighted females over multiple years to assess various parameters, such as age at first reproduction, birth season, birth intervals, and productivity. The description of reproductive patterns over time is possible using such data, which enables investigations on the variation in reproductive rates within threatened areas, where dolphin studies are often lacking.
In recent years, a reduction of 30% was recorded in the number of bottlenose dolphins inhabiting the Mediterranean Sea (Bearzi et al. 2009), largely as a consequence of deliberate killing, habitat degradation, and overfishing (Reeves & Notarbartolo di Sciara 2006). This suggests that the reproductive output of bottlenose dolphins might be largely influenced by the nature and extent of these threats in different regions (Bearzi et al. 2008, Reeves & Notarbartolo di Sciara 2006, Gnone et al. 2011. Assessing female reproduction in Mediterranean bottlenose dolphins is of paramount importance to identify the viability of the population and formulate adequate conservation measures (Caswell 2001, Fujiwara & Caswell 2001. A 13 yr (2005−2017) photo-ID study of bottlenose dolphins in the Aeolian Archipelago, in the Southern Tyrrhenian Sea (Italy), has revealed complex societies and behaviours adapted to both environmental and anthropogenic conditions (Blasi & Boitani 2012, Blasi & Boitani 2014. The encounter rate of dolphins has decreased significantly in recent years, with data suggesting that Aeolian bottlenose dolphins have high site fidelity to the area (Blasi & Boitani 2012, Leone et al. 2019. Beginning in spring−summer, females show strong associations with their calves, spending extensive time in the safest areas for feeding, socializing, resting, and caring for their calves (Blasi & Boitani 2012, Blasi et al. 2015. In contrast, males adapt their behaviour to achieve the highest probability of locating and capturing prey in fishing areas (Blasi & Boitani 2012, Blasi et al. 2015. In particular, some males have specialized to aggregate in small groups to depredate gillnets (Blasi & Boitani 2012, Blasi et al. 2015. Overfishing is high, and a marine protected area has not yet been implemented in the Aeolian Archipelago, suggesting that food resources might be scarce for the dolphins (Blasi & Boitani 2012, Blasi et al. 2015. A recent paper also showed that the Aeolian bottlenose dolphins are strongly malnourished and present skin lesions, such as gunshot wounds and mutilations (Leone et al. 2019).
In this study, data from a long-term mark−recapture study were used to document the reproductive history of female bottlenose dolphins living in the Aeolian Archipelago. Using high-quality dorsal fin photographs, the reproductive patterns of individual females and their calves were tracked to estimate the birth period, calf productivity, interbirth intervals, age of first reproduction, calf mortality, weaning period, group size, mother−calf postweaning associations, and association patterns among females.

Survey and photo-ID data
The Aeolian Archipelago is located in the Southern Tyrrhenian Sea, north of Sicily, Italy (38°35' N, 14°34' E). Dedicated boat surveys were performed from June to September throughout 2005−2017 using random transects within a study area of 400 km 2 (Blasi & Boitani 2012, Blasi et al. 2015, Leone et al. 2019. Bottlenose dolphins were observed using a combination of focal group observations (Shane 1990), with instantaneous data sampling (Altmann 1974), and photo-ID techniques (Würsig & Würsig 1977, 1979, Würsig & Jefferson 1990. Assuming that clustered animals were also interacting (Whitehead & Dufault 1999), associations were defined by membership to the same group, i.e. animals photographed in the same group and moving in the same general direction, interacting or engaged in similar activities, were considered to be associated (Shane et al. 1986). The focal group was defined as individuals within 10 m of at least 1 other dolphin in the group. When more than 1 focal group was encountered in the same survey, each sub-group was recorded as an independent sighting (Mann 1999). In the event that a group split up, 1 of the sub-groups was followed at random and independent of group size or activity (Mann 1999).
Each individual was photographed by standard photo-ID techniques. When possible, videos taken during each sighting were used for sex identification (Blasi & Boitani 2014, Blasi et al. 2015, and individuals were identified by the natural markings on the body, especially on the dorsal fin. High-resolution photographs of distinctive dorsal fins were used to match individuals with a photo-ID catalogue of known individuals (Würsig & Würsig 1977). Each dolphin in the catalogue was classified with a code (PHD) followed by a number (n) according to its photo-identification date (PHDn). In addition, each calf was classified with the code of the mother followed by the number of offspring produced by the female since the start of the study (i.e. PHDn[m]). For example, the third calf of female PHD9 was classified as PHD9[3], while PHD9[1][1] is the calf of calf PHD9[1] who became a mother in the study period. The mean (± SD) number of calves for reproductive females was also estimated.

Age and sex
Each year, all individuals were classified according to their relative size assessed through photo-ID (Blasi & Boitani 2012, Leone et al. 2019: (1) adults: large and robust animals, with dark skin colour and many marks on the dorsal fin and body, and often accompanied by a calf; (2) juveniles: less robust individuals at least two-thirds the length of adults, usually with less distinctive nicks, or without nicks in their dorsal fins and not associated with an adult; and (3) calves: one-third the size of adults, in close association with an adult and usually without nicks in their dorsal fins. Calves were also identified according to typical surfacing behaviour, swimming in infant position, in contact underneath the mother, or spending all of their time with an adult (Grellier et al. 2003).
Dolphin sex was determined by opportunistic viewing of the genital region and later confirmed by pictures and frame shots of the genital slit taken during each sighting. In this study, it was possible to collect several pictures of each individual, which helped in sex determination. Adult bottlenose dolphins lack obvious and detectable sexual dimorphism (Read et al. 1993), and it is not always easy to obtain good quality pictures of the genital area for sex determination . Reproductively mature females (Grellier 2000) were identified when they maintained a close and lasting (> 2 yr, and in all sightings for each year) relationship with a calf, which was presumed to be their offspring. Based on long-term photo-ID data, all calves analysed in our study were strongly associated with known females identified by their dorsal fin. It is well known that alloparental care could lead to false mother−calf pairings in bottlenose dolphins (Whitehead 1996), but this potential error was re-moved by ensuring at least 2 photographic frames from the same sighting and more than 2 frames within the same season showing the calf in echelon position with the presumed mother.

Occurrence and group size
The occurrence (%) for years (number of female sightings/years of sightings) and the occurrence (%) for sightings (number of female sightings/total number of sightings) were calculated for each individual.
The group size of females was visually determined in situ and later verified by standard photo-ID techniques and videos taken during each sighting. The mean (± SD) group size was calculated for each female on pooled sightings for the entire study period (Blasi & Boitani 2014). A linear regression analysis (PAST 4 Software version 4.03) (Hammer et al. 2001) was used to examine the correlation between mean group size and the number of calves for females.

Birth period
The birth period was estimated based on (1) the time between the last sighting of the mother before giving birth and the sighting of the mother with a calf and (2) the proportional body length between mother and calf (i.e. calves less than 1 yr old were about onethird of the length of the mother). The birth period was estimated to within 1 mo for all calves up to 2 yr of age sighted before the end of the weaning period.

Interbirth interval
The years between the estimated birth period of a surviving calf and a subsequent birth were used to determine the interbirth intervals for each female dolphin. In order to reduce the probability that the birth of a new calf was missed for the less frequently sighted females, and the chance of the interbirth interval being shorter if mothers lost unsighted calves, the interbirth interval was calculated only for females that were adults in the first year of the project and were sighted for at least 9 years within the study period and more than 5 times each year. Finally, only females with a surviving calf of known birth within 1 yr, which did not disappear during the interbirth interval for more than 1 yr from the study area, were included in the interbirth interval analysis.

Weaning period
The weaning period was estimated based on the time between the birth period and the last sighting of the calf nursing (i.e. beak to the mammary slit). Since not all calves were observed nursing during the study period, the weaning period was also calculated based on the last sighting of the calf in infant position (Mann & Smuts 1998). If the year of birth could not be determined, the calf was not included in the age-specific weaning period estimation. When a new calf was born, the weaning period for the old calf was estimated based on the last sighting at which the calf stopped its exclusive association with the new mother−calf pair (but not necessarily in infant position) and with the maternal group during the same year/season. A weaning period of < 6 yr was fixed to provide a threshold value of assurance that the old calf had stopped nursing following the birth of a new one , Foroughirad & Mann 2013, Karniski et al. 2018).

Mother−calf associations postweaning
Postweaning mother−calf associations were estimated from the last sighting at which a calf was exclusively with the maternal group until it was sighted in other dolphin aggregations, in the same year/season. It was possible to keep track of juveniles/adults only for calves that had distinctive dorsal fin markings before the end of the weaning period. In particular, postweaning mother−calf associations were estimated only for calves that were confirmed to be out of the weaning period, i.e. they were no longer nursing, swimming in infant position, or showing exclusive mother−new calf associations. For the estimation of the postweaning period, the sample size varied according to known birthdates, weaning dates, and survivorship of the calves. Furthermore, calves with an unknown birth period within 1 yr were not included in the calculation of mother−calf postweaning associations. For each calf, the residence time with the mother was calculated as the sum of the weaning period and the postweaning period.

Calf mortality
The death by calf age (calf mortality) was also estimated for each adult female. In order to reduce the probability that the death of a calf was not recorded as a result of rare sightings of the mother, calf mortality was calculated only for females that were observed for at least 9 years within the study period, more than 5 times each year, and had not disappeared from the study area for more than 1 yr. Moreover, only calves with a known birth period within 1 yr were included in calf mortality analyses to ensure that unsighted calves had been weaned rather than presumed to have died. Specifically, for calves less than 2 yr old (i.e. an arbitrary time to assure exclusive mother−calf associations within the weaning period), if mothers were sighted repeatedly without the calf during the same year/season, it was assumed that the calf had died. Calf mortality was also estimated postweaning (i.e. more than 2 yr) for calves that had distinctive dorsal fins before the end of the weaning period, using the last date on which the calf was sighted in the study area (i.e. disappearing from the study area for more than 2 consecutive years).

Female−female association patterns
The half-weight index (HWI) was calculated by the equation 2N/(N PHD1 + N PHD2 ), where N is the number of times that individuals PHD1 and PHD2 were seen together, and N PHD1 and N PHD2 are the total number of sightings of each individual (Cairns & Schwager 1987, Bräger et al. 1994, Ansmann et al. 2012. To ensure the independence of sampling and avoid serial autocorrelation of sightings, only the first sighting was used when an individual dolphin was sighted more than once in a day (Bëjder et al. 1998). Pooled HWIs (years combined from 2005 to 2017) were calculated for pairs of females, juveniles, and calves that were well-marked and recaptured more than 6 times during the study period. Only calves sighted for more than 3 consecutive years and with complete known histories were included in the analysis due to the bias of their unique dependent relationships with their mothers. Group membership samples were used to obtain an association matrix of all individuals (Whitehead & Dufault 1999). A PCA was applied to the HWIs to investigate whether association patterns differed significantly among females and their calves. Independent principal components (factors) were extracted from the original variables, from which the leading ones (i.e. those which explained more than 80% of variance) were plotted to identify significant association patterns (Blasi & Boitani 2014). Finally, a simple linear regression analysis was used to investigate the correlation be tween the HWIs and the number of calves for each female (Blasi & Boitani 2014). PAST 4 Software (version 4.03) was used to examine these relationships (Hammer e al. 2001).

RESULTS
We photo-identified 42 dolphins by the permanent marks on their dorsal fins during 272 focal follows between 2005 and 2017, i.e. 19 adult dolphins at their first sighting and 24 adults at their last sighting. Analysis of photographs of the genital area definitively identified 8 females and 14 males. By the end of the study, 80% of adults were classified as having high site fidelity, including 8 females, 14 males, and 2 individuals of unidentified sex. These dolphins were sighted at least 9 years of the study period, more than 5 times in the same year, and they never disappeared from the study area for more than 1 yr (Table 1).
Only 7 females were clearly associated with a calf during the study period (Table 1), and a total of 16 calves were photo-identified in strong association with the mother ( Table 2). The number of calf sightings varied from 2 to 57 within the study period ( Table 2). The mean number of calves for reproductive females was 2.3 ± 1.1 (SD) and varied from 1 to 4 calves per female during the study period (Table 1). PHD4 was the most reproductive female with a total of 4 calves, followed by PHD9 and PHD10 with 3 calves each (Table 1). Four females (PHD3, PHD4, PHD9, and PHD10; 57.1% of females) disappeared from the study area 1 yr before having a calf, to return the following year with a newborn (Fig. 1). However, 2 females were not sighted in the last years of the project (Fig. 1).

Occurrence and group size
Females PHD3 and PHD4 were observed more frequently than other females in the study area (Table 1). The mean group size of females (Table 1) varied from 4.5 ± 2.9 (SD) for PHD21 (88.8%) to 6.8 ± 0.8 for PHD9[1] (89.3%). Furthermore, the number of calves appeared to be influenced by the mean group size of the mother, with females sighted in larger groups producing more calves (n = 6; R 2 = 0.7, p < 0.05) ( Table 1).

Birth period
Our sample for the estimation of the birth period included birth age estimates accurate to within 1 mo for 6 calves (37.5%), within 1 yr for 8 calves (50%), and within 2 yr for 2 calves (12.5%) ( Table 2). The birth period was seasonal (spring−summer), peaking in June/July (Table 2).

Interbirth interval
One of the females (PHD6) sighted for at least 9 years, more than 5 times per year, and identified as an adult in the first year of the study was sighted with only 1 calf during the study period and, consequently, it was not included in the interbirth interval calculation (Fig. 1). Two females, PHD3 and PHD5, had a time gap between the weaning of a calf and the birth of another, such that pregnancy and/or loss could have occurred, while the others had a weaning period that overlapped with the birth of the new calf ( Fig. 1) was calculated for 5 females and was 3.5 ± 1.6 yr (SD), ranging from 2 (for PHD4) to 7 yr (for PHD3), although it was observed to be between 2 and 4 yr for 4 females.

Weaning period
The weaning period was calculated based on 8 interbirth intervals for 11 calves with a known birth date within 1 yr, while for the others a range period was established (Fig. 1). The mean weaning age was 3.23 ± 0.61 yr (SD), ranging from 2 to 4 yr, but 1 calf (PHD5[1]) may have been weaned as late as its fifth birthday.

Postweaning associations
The postweaning period was estimated for 9 calves which had clear permanent marks before the end of the weaning period, allowing us to follow the complete history of each individual (Fig. 1) [2]), as a result of premature death or migration to other areas, or in cases when they were still juveniles at the end of the project. From the postweaning period, the time of residence in the maternal group was estimated to range from 4 (for PHD4[1]) to ≥ 10 yr (for PHD9[1]), but it was observed to be ≥ 6 yr for 5 calves (Fig. 1).

Age of first reproduction
We were able to document the reproductive trend of female PHD9[1] in detail. This individual was focally followed from its birth to adult age, and its first pregnancy was recorded at the age of 7.5 yr ( Table 1). The ages of first reproduction of the remaining 8 calves which had clear permanent marks before the end of the weaning period remain unknown.

Calf mortality
The number of calves which survived to weaning age (< 6 yr) over the 13 yr of the study was also estimated for each female, resulting in 12.5% mortality by age 1 (2 calves

Female association pattern
The score plot of the main principal components (Factors 1 and 2) calculated for the HWI showed the association pattern among females and calves/ juveniles with known/unknown mothers (Fig. 2). Two well-distinguished clusters were identified in the plot (Fig. 2): Cluster 1 included 3 females, and Cluster 2 included 5 females. However, the number of calves was not significantly related to the HWI of the mother (R 2 = 0.07, p > 0.05) ( Table 1).

DISCUSSION
Reproduction of bottlenose dolphins in their natural environment has been extensively investigated in the scientific literature (Wells 2000, Steiner & Bossley 2008, Rossi et al. 2017, showing a reproductive trend that may consistently vary among different populations resident in the study area (Haase & Schneider 2001, Henderson et al. 2014, Fruet et al. 2015 and other populations (Brough et al. 2016, Cheney et al. 2019. However, very few data are available on the reproductive output of Mediterranean bottlenose dolphin populations, where flexible social organization, op portunistic diet, and behaviour may allow them to endure some of the effects of overfishing and habitat degradation (Bearzi et al. 2009, Gnone et al. 2011. In this study, long-term photo-ID data (Blasi & Boitani 2014, Blasi et al. 2015, Leone et al. 2019 allowed the estimation of important reproductive parameters in bottlenose dolphins living in one Mediterranean area.

Female productivity, group size, and association patterns
In this study, only a few females were clearly associated with 1 or more calves during the study period and consequently were classified as reproductive. Moreover, the population could not be maintained by the 16 calves born in 7 years. The number of calves per female was highly variable, ranging from 1 to 4, and suggested that reproductive success might be related to several factors, as already found in other bottlenose dolphin populations (Kellar et al. 2017, Robinson et al. 2017, Baker et al. 2018). Many studies , showing the association patterns among female bottlenose dolphins (in bold) and juveniles/calves (plain text) with the known mother. Three more juveniles with unknown mothers (PHD13, PHD14, PHD24) were also included to identify potential correlations with other dolphins. For each individual, the ID code and the years of first and last sighting are also indicated in the plot. Grey text indicates calves/juveniles with known mothers that were not included in the analysis, as they were sighted fewer than 6 times have shown that female bottlenose dolphins may play an important role in the formation and maintenance of the community social arrangement and conservation of the population , Blasi & Boitani 2014, Krzyszczyk et al. 2017. Our results have shown that females have significantly more preferred associates and favour larger groups compared with males (Blasi & Boitani 2012, Blasi et al. 2015. Groups of bottlenose dolphins with calves were previously found to be significantly larger than those without calves (Hubard et al. 2004, Blasi & Boitani 2012, and in this study the number of calves produced was influenced by the mean maternal group size. The trend of finding calves in larger groups suggests that larger group size may benefit mother−calf pairs (Schaffar-Delaney 2004, Gibson & Mann 2008, Robinson et al. 2017; however, in this study, the number of calves per female was not significantly related to the strength of the associations of the mother. In Sarasota Bay, Florida (USA), Wells (1991a) showed that survivorship of bottlenose dolphin calves was related to group size. The functions of dolphin groups are numerous , and larger aggregations could be advantageous for female bottlenose dolphins and their calves in many different ways, including for the protection from predators or conspecifics (Corkeron et al. 1987, Cockcroft et al. 1989) and defence against sexual coercion and infanticide by males (Hrdy 1979, Patterson et al. 1998, Fearnbach et al. 2012. Living in large groups may de crease the risk of predation through increased vi gilance (Evans 1987, Alcock 1998, Fellner 2000. Moreover, larger groups may also benefit mother− calf pairs by increasing foraging efficiency, and consequently fertility and productivity, as the location of food resources is facilitated by collaboration among individuals through increased alertness and surveillance. Finally, larger groups may provide increased predation ability , Fellner 2000 and more opportunity for social learning (Norris & Dohl 1980, Johnson & Norris 1986, Wells 1991b. As a result of the high vulnerability of newborn calves and the high energy costs incurred by lactating females, the different advantages of group living appear to be of primary importance for the survival of calves and are likely to explain why newborn dolphins are found in larger groups. We found that some females disappeared from the study area before giving birth to a new calf. The time interval when a female leaves the study area after giving birth can be used as an index to understand if the individual considers the habitat safe for rearing the calf. In Shark Bay (Australia), Smolker et al. (1992) noted that after the conception of calves, female bottlenose dolphins remained in their natal area and continued to associate with their mothers. In addition, with the birth of their first calf, females in Sarasota Bay returned to their maternal group (Wells et al. 1987). In our study, some females may have left the study area temporarily to give birth in safer areas (Wells et al. 1987). Fluctuations in the annual residence of females with calves returning to the study area after a segregation period suggest that females may temporarily separate from other females.

Birth period
It is already known that female bottlenose dolphins may coordinate births and, in the Aeolian Archipelago, the birth period takes place during summer, particularly from June to July. Warm water might influence birth seasonality, since it is known to be thermally more efficient for both the calf and the mother . Firstly, warm water results in greater food availability, which favours seasonal births. Consequently, the food intake of females can be maximized during mid-lactation when nutritional stress is likely to be greatest (Oftedal 1997). Secondly, during the warmest months, the energy de mand on lactating females is reduced and less energy is required by the newborn to regulate body temperature (Wells et al. 1987). Seasonality as found in our investigation has also been documented in other studies (Urian et al. 1996, Thayer et al. 2003, Henderson et al. 2014. For example, the calving period of bottlenose dolphins from the southwestern Atlantic Ocean is associated with warmer surface waters (Fruet et al. 2015). In contrast, seasonality was not found in Portugal, where births are registered year-round (Dos Santos & Lacerda 1987), and a relationship was not found in populations off Florida and Texas (Urian et al. 1996). Local environmental conditions are thought to have an important influence on the seasonality of reproduction (Urian et al. 1996, Tezanos Pinto 2009).

First reproduction
The age of first pregnancy was recorded to be 7.5 yr, which is within the ranges reported for other bottlenose dolphin populations (Fruet et al. 2015, Robinson et al. 2017, Kemper et al. 2019. In captivity, a bottlenose dolphin female typically begins to be reproductive at 7−10 yr, but may also begin earlier, between 5 and 7 yr (Schroeder 1990). In Sarasota (Wells et al. 1987), multiple births were recorded by females between 6 and 7 yr old, in populations where sexual maturity is reached at about 5−12 yr (Tolley et al. 1995). Along the central Atlantic coast, females become sexually mature between 7 and 13 yr (Mead & Potter 1990). On the other hand, the average sexual maturity of females in South African waters is reached around 9−11 yr (Wells & Scott 1999). During the first reproduction, females are considered to be young, and consequently, the first calves recorded (Mitcheson 2008) are considered to be more vulnerable to predators. Moreover, young females may not be large or fat enough to be able to nurse a growing calf . However, in this study, not enough data were available to establish if first-time mothers are more likely to lose their first calf than more experienced mothers (Clutton-Brock 1988).

Interbirth interval
The interbirth interval in this study was most commonly between 2 and 7 yr, with several females having intervals between 2 and 4 yr. Variability in interbirth intervals can be potentially attributed to several factors, including female age and body mass, year of first reproduction, previous number of calves, calf loss, and weaning period , Mitcheson 2008, Fruet et al. 2015. Few studies have reported an interbirth interval of < 3 yr for a surviving calf; one study reported an interbirth interval of 3 yr for bottlenose dolphins in Brazil (Fruet et al. 2015), whereas intervals of about 3−5 yr were found to be more common, e.g. in bottlenose dolphins from Sarasota, which had an interbirth interval of 4 yr ). The mean interbirth interval of bottlenose dolphins was 5.3 yr in Doubtful Sound, New Zealand (Henderson et al. 2014), 3.8 yr in Scotland (Mitcheson 2008), and 4.1 yr in Shark Bay, Australia .

Calf mortality
In our study, calf mortality by age 1 was particularly high (12.5%), and 2 other juveniles also disappeared 1 yr or more after weaning and were presumed dead. There may be some underestimation of mortality in young calves, given that some calves likely died before being sighted (Tezanos-Pinto et al. 2015). As our data derive from multiple sightings of mother−calf pairs during the weaning period, we cannot exclude a potentially high level of uncertainty in the determination of whether a calf survived, was successfully weaned and became independent from its mother, or whether a calf died, and a new calf was born before the end of the calving period. The death of a calf in the first year can lower the mean calving interval (Perrin & Reilly 1984, Cockcroft & Ross 1990, Bearzi et al. 1997, re sulting in a lower mean interbirth interval. The greatest mortality risk occurs within the first months , which may lead to the mother becoming pregnant again in the year immediately following the death of a juvenile. Furthermore, a calf was assumed to have died if the mother was re sighted without the calf upon 2 consecutive encounters (Steiner & Bossley 2008). The causes of calf mortality are largely unknown in this area. Shark attacks and infanticide are rare in the Mediterranean (Bianucci et al. 2002), but possible causes of death could be associated with changes in diet, age, or malnutrition of the mother due to the low concentrations of prey in the area , Blasi & Boitani 2014, Rossi et al. 2017. In Shark Bay, foraging success by the mother is a determining factor in the survival of the calf . If malnutrition due to insufficient food resources is not the direct cause of calf mortality, other factors likely affect calf survival, including lesions from fishing gear or competition with the artisanal fishery (Blasi & Boitani 2012, Blasi & Boitani 2014.

Weaning period
Findings from our study suggest that weaning can last between 2 and 4 yr in bottlenose dolphins from the Aeolian Archipelago, with a mean weaning age of 3.23 yr; however, 1 calf may have been weaned as late as its fifth birthday, due to its exclusive association with the mother when a new calf was born. Although nursing or swimming in infant position are more reliable indicators for the estimation of the weaning period, an exclusive association pattern with the mother can also be used . Accordingly, our data show that calves nurse, at least occasionally, for the entire period of infant position swimming, typically 2−4 yr, and occasionally up to 5 yr. These periods concur with previous estimates of minimum weaning age at 18−20 mo in other bottlenose dolphin populations (Smolker et al. 1992, Wells & Scott 1999. As the separation between mother and calf tends to occur before the birth of the next calf , Wells 2000, Fruet et al. 2015, the mean interbirth interval may at least roughly coincide with the time of weaning in these bottlenose dolphin populations.

Postweaning associations
In bottlenose dolphin populations, mother−calf relationships may extend beyond weaning, and calves are likely to remain with their mothers for any length of time (Shane et al. 1986, Bearzi et al. 1997, Grellier et al. 2003. In this study, it was possible to keep track of calves (until they reached juvenile/ adult stages) that had clear permanent marks on the dorsal fin before the end of the weaning period. For these dolphins, the postweaning period lasted from 4− 10 yr, but it was more than 6 yr for 5 calves. These long periods of dependency on the mother stress the importance of mother−calf relationships for the development of calves in this population (Grellier et al. 2003). Few studies provide information on postweaning associations in bottlenose dolphins, especially in the Mediterranean Sea, but there is strong evidence that mother−calf associations lasting several years may be related to the benefits derived from these associations (Van Schaik 1989, Gibson & Mann 2008, Krzyszczyk et. al 2017, such as for foraging or de fence from predators/conspecifics (Krzyszczyk et. al 2017). After birth, the early interactions with mothers teach calves important social/ behavioural skills, which will then be extended to a wider social context involving different participants (Makecha 2005). During that time, mothers will increase the chances of survival of their calves by transmitting vital information including food distribution, foraging strategies, predator avoidance, recognition of threats, and patterns of social interactions (Wells et al. 1987, Grellier et al. 2003. These social aspects of calf development are as necessary as motor coordination to gain independence (Wells et al. 1987).

CONCLUSIONS
Our results provide important information on female reproductive output in a Mediterranean bottlenose dolphin population. These data and previous results highlight the need to implement urgent conservation actions to protect this threatened bottlenose dolphin population. The incorporation of these findings into population viability analyses could help in the design of conservation initiatives, thereby con-tributing to the growth of this population. The need for a larger data set is important for future investigations to understand the primary mechanisms underlying the development and maintenance of female− calf associations. Of particular interest will be to determine if the low reproductive output is due to calf mortality and/or to female condition, such as fertility and/or malnutrition.