Patoku Lagoon Field Station

Researchers were permanently based at Patoku (lat 41°38′N, long 019°35′E)—a lagoon system that links to Drini Bay—from June to September each summer (Fig. 1). We worked closely with artisanal fishermen, obtaining sea turtles that were caught incidentally as bycatch (nontarget species).

View Figure

• Download as Powerpoint
• Download Figure

Drini Bay

Drini Bay extends about 30 km from Buna River, at the northern border with Montenegro, to Cape Rodoni, an 8-km peninsula into the Adriatic Sea. The bay is sparsely populated, especially in the southern half; road access was very difficult but is now being improved. The seabed is mostly sand and mud, with high terrigenous inputs entering the bay from 4 major rivers (Buna, Drini, Mati, and Ishmi). Benthic fauna is abundant, marine flora are sparse, and subsurface visibility is usually 0. The main economic activity is artisanal fishing, usually conducted from small flat-bottomed rowboats. Additionally, there is a small trawling fleet based at the port of Shengjin in the northeasternmost part of the bay that frequently fishes illegally within a few meters from the shore; law enforcement resources are extremely limited.

Stavnikes

Stavnike fish traps (pound nets) are large rectangular structures (about 90 m long and 8–10 m wide) that consist of a series of chambers, formed by nets attached to tall wooden posts, erected in water depths of 6–8 m and with a long barrier net that extends to the shore (see White et al. 2010a). Sea turtles often entered the stavnikes, some being recaptured on several occasions. They can swim around freely within the enclosure, feed on fishes or other prey, and, crucially, surface to breathe; therefore, stavnikes appear to have a very low impact on them (White et al. 2010a, 2011a). Fish traps are usually erected during April–May and operate until either the weather and heavy seas force them to be dismantled (this varied annually but was between July and September) or the fish catch was reduced to the point that it was not worth the daily effort. Two stavnikes were monitored closely: 1) near the outflow of River Mati, about 150 m offshore, and 2) northwest of River Ishmi outflow, about 1800 m offshore; both groups of fishermen were based at Patoku. Between September and April each year, these fishermen catch eels (Anguilla anguilla) in the very shallow Patoku lagoons instead, as their flat-bottomed boats are unsafe in heavy seas.

Sea Turtle Morphometrics

The curved carapace length (CCL) and curved carapace width (CCW) were measured (Eckert et al. 1999), and turtles were allocated into 10-cm size classes (length–frequency distribution) based on their CCL. Occasionally, measurements were incomplete because of injuries (e.g., part of carapace missing).

Laparoscopy was unavailable, so the tail morphology was used instead to determine if turtles were adult or still immature (i.e., the developmental stage of secondary sexual characteristics). Three measurements were recorded from the tail ventrally (Casale et al. 2005; White et al. 2008, 2010a, 2010b, 2011a): distance from the posterior margin of the plastron to the midline of the cloacal opening (plas-clo), total tail length (TTL), and distance from the tip of the tail to the posterior margin of the carapace (± carapace): 1) the rearmost margin of the carapace was taken as the 0 datum (0.0); 2) postcarapace (> 0.0), the more developed tail in males extends beyond the carapace margin; and 3) subcarapace (≤ 0.0), females and smaller juveniles of both sexes tend to have shorter tails.

Loggerhead turtles were separated into 2 groups: long-tailed (the tail protruded beyond the supracaudal scutes) and short-tailed (the tail tip did not extend beyond the carapace margin). One-way analysis of variance (ANOVA) was used to compare morphometric values between the 2 groups within each 10-cm size class.

Recaptures and Site Fidelity

Several previously tagged-and-released loggerheads were recaught in subsequent years, allowing incremental growth rates to be calculated; these provided limited insight into the onset of maturation and size at maturity. Some turtles were caught more than once within the same year, suggesting that those animals remained locally during the summer months, presumably for foraging. Incremental growth rates were calculated using the following method: (measurement at recapture [cm CCLr] − measurement at tagging [cm CCLt])/interval between tagging and recapture in years (e.g., 14 mo  =  1.17 yrs).

Male Turtles

The bycatch each year at Drini Bay included an unusually high proportion (27%) of male turtles; sexual categorization was based on tail morphology (White et al. 2008, 2010a, 2010b, 2011a, 2011b). In 2008, we showed that there were highly significant differences in the 3 tail measurements between 14 immature and 4 adult male C. caretta (White et al. 2008). By the study's end in 2010, we had tagged 35 adult males: most were in the 70-cm size class (n  =  24); 2 were just less than 70 cm (CCLs  =  68.5 and 69.0 cm); the remaining 9 were over 80 cm CCL.

For male turtles showing definite tail development but not yet mature, the most populous size class was 60 cm (n  =  39); 7 were in the 50-cm size class (the smallest turtle showing definite tail development had a CCL  =  52.0 cm), but there were also 29 subadults in the 70-cm size class and 1 in the 80-cm class (CCL  =  81.0 cm).

Once again, we found highly significant differences between the 3 tail measurements of adults and developing males (White et al. 2010b, 2011b). There was a 12.5-cm size range overlap between the smallest adult (CCL  =  68.5 cm) and the largest subadult (CCL  =  81.0 cm; its tail was nearer in size to an adult's than a juvenile's but still not mature).

Short-Tailed Turtles

The majority of loggerheads (n  =  259) had shorter tails; that is, the tail tip did not extend beyond the posterior margin of the carapace, meaning that they were probably females or smaller juveniles of either sex. Based on our knowledge of size classes obtained from developing and adult males, we allocated all turtles within each 10-cm CCL size class into 2 groups, long-tailed (postcarapace) or short-tailed (subcarapace; Table 2), and then compared these using 1-way ANOVA.

Table 2. Caretta caretta. Morphometric data for curved carapace length (CCL) size classes. Turtles were grouped by tail length (postcarapace or subcarapace) within each size class. All turtles in the 2 smallest size classes were short-tailed. CCW  =  curved carapace width; plas-clo  =  distance from plastron to midline of cloaca; TTL  =  total tail length; ± carapace  =  tail is longer than or shorter than the carapace rear margin.

View Fullscreen

For the 50–80-cm size classes, ANOVAs showed no significant differences between long- and short-tailed individuals for CCL or CCW (p ≥ 0.5), but there were highly significant differences in all 3 tail measurements (plas-clo, TTL, and ± carapace) between the long-tailed and short-tailed groups in each class: 80-cm size class: n  =  10 long-tailed, 4 short-tailed (plas-clo: F_1,12  =  37.36, p < 0.01; TTL: F_1,12  =  45.01, p < 0.01; ± carapace: F_1,12  =  45.19, p < 0.01); 70-cm size class: n  =  53 long-tailed, 49 short-tailed (plas-clo: F_1,100  =  103.45, p < 0.01; TTL: F_1,100  =  112.78, p < 0.01; ± carapace: F_1,94  =  180.17, p < 0.01; 4 postcarapace and 2 subcarapace records were data deficient); 60-cm size class: n  =  41 long-tailed, 108 short-tailed (plas-clo: F_1,147  =  93.13, p < 0.01; TTL: F_1,147  =  130.90, p < 0.01; ± carapace: F_1,143  =  317.75, p < 0.01; 3 postcarapace and 1 subcarapace records were data deficient); 50-cm size class: n  =  7 long-tailed, 78 short-tailed (plas-clo: F_1,83  =  7.44, p < 0.01; TTL: F_1,83  =  11.77, p < 0.01; ± carapace: F_1,81  =  42.00, p < 0.01; 2 postcarapace records were data deficient).

In the smallest size classes (40 cm; n  =  17; 30 cm; n  =  3), all turtles were short-tailed; there were no postcarapace caudal data, and ANOVA was not used.

Incremental Growth, Carapace Data

CCL and CCW measurements for 26 turtles recaptured in subsequent years were compared with the initial data (at tagging); any growth increases were related to the interval between captures. CCL growth rates for 26 loggerheads ranged between 0.0 and 4.9 cm/yr (mean  =  1.7; SD  =  ± 1.4). CCW growth rates ranged between 0.0 and 2.8 cm/yr (mean  =  1.1; SD  =  ± 0.8). The longest interval between captures was 84 mo, and the shortest was 9 mo (mean interval  =  23; SD  =  ± 19.4).

Incremental Growth and Tail Data

There were only 4 records for tail-morphological data at release and recapture; these data were not collected prior to the present study, and 6 recapture records were excluded (data deficient) because researchers, outside the field season, deviated from the data collection methodology.

Table 3 shows the tail measurements for 4 recaptured loggerhead turtles; the intercapture interval varied between 9 and 23 mo. The 2 most important animals are PA040 (tag no. AL0175) and PA233 (tag no. AL0125); both are males that began maturation during the intercapture period, noted by the increased TTL and postcarapace measurements. PA075 (tag no. AL0135) may prove to be a female, as the tail is still subcarapace, but the CCL grew rapidly (+ 3.0 cm in 13 mo). PA037 (tag no. AL0065) is a juvenile and could be either male or female; its tail measurements are not yet conclusive, but as the postcarapace[r] is now 0.0, this may be a male; eventual future recapture will provide more evidence.

Table 3. Caretta caretta. Growth increments are given for tail measurements in 4 loggerhead turtles. Interval is period between captures in years (i.e., number of months/12). CCL  =  curved carapace length; [t]  =  measurement at tagging; [r]  =  measurement at recapture; increment is the growth; rate is growth/interval (cm/yr); plas-clo  =  posterior margin of plastron to cloaca; TTL  =  total tail length; postcara is the tip of tail to posterior margin of carapace (minus sign denotes subcarapace); DD  =  data-deficient record.

View Fullscreen

The measurements for the 4 females with the longest tails were as follows: 1) CCL  =  80.0 cm, plas-clo  =  17.5 cm, TTL  =  21.0 cm, subcarapace  =  0.0 cm; 2) CCL  =  79.0 cm, plas-clo  =  17.5 cm, TTL  =  22.0 cm, postcarapace  =  0.5 cm; 3) CCL  =  76.0 cm, plas-clo  =  17.0 cm, TTL  =  21.0 cm, subcarapace  =  −1.0 cm; and 4) CCL  =  76.0 cm, plas-clo  =  12.0 cm, TTL  =  20.0 cm, subcarapace  =  −2.0 cm.

Marine Habitat Use

White's (2007) earlier study at Lampedusa, Italy, showed that the CCL size classes (mean CCL  =  52.4 cm; SD  =  ± 13.1 cm; range  = 19.0–85.0 cm; n  =  176) for loggerhead bycatch in the southern Ionian Sea were typically a size class smaller than those reported here for Drini Bay.

Table 4 compares the CCL size classes for bycaught turtles at Lampedusa and Drini Bay. At Lampedusa, the most populated size classes were 40–60 cm; at Drini Bay, 50–70 cm. Lampedusan bycatch included turtles in the 10–20-cm size classes; these were absent at Drini Bay. There were no turtles larger than the 80-cm size class from either location.

Table 4. Caretta caretta. Curved carapace length (CCL) was used to allocate turtles into 10-cm size classes. Turtles were captured in fisheries at 2 sites: Lampedusa (Italy) and Drini Bay (Albania). At Lampedusa, the majority of turtles were in 40–60-cm size classes; at Drini, 50–70-cm size classes. The animal in the 10-cm size class was a loggerhead hatchling found at sea near to Lampedusa (this was not tagged).

View Fullscreen

Green Turtles (C. mydas)

Green turtles were occasionally captured (bycatch): 4 in Drini Bay stavnikes (1 in 2008; 3 in 2010), and in 2009 another C. mydas was captured in a net at Orikum near Vlore (Fig. 1). Morphometric data (cm) were as follows: mean CCL  =  46.9 (SD  =  ± 18.0; range  =  30.0–67.0; n  =  5), mean CCW  =  43.2 (SD  =  ± 17.7; range  =  28.0–62.5; n  =  5), mean plas-clo  =  7.3 (SD  =  ± 3.2; range  =  4.0–11.0; n  =  5), mean TTL  =  10.2 (SD  =  ± 4.4; range  =  6.5–15.0; n  =  5), mean postcarapace  =  1.8 (SD  =  ± 0.3; range  =  1.5–2.0; n  =  3), and mean subcarapace  =  −0.3 (SD  =  ± 0.4; range  =  −0.5–0.0; n  =  2). Two of the green turtles were sexually developing males (tag nos. AL0224/AL0223 and AL0264/AL0265).

Male Turtles

The marine ecology and distribution of sea turtles, especially for males, has received less research attention than the substantial nesting beach studies globally (Bjorndal 1999; Schroeder et al. 2003). For instance, Ehrhart (1982) found just 14 references concerning male sea turtles in his literature review at that time, although this has now improved (Marquez 1990; Bolten and Witherington 2003; Ehrhart et al. 2003; Limpus and Limpus 2003; Margaritoulis et al. 2003). In the Mediterranean, Margaritoulis (1982) reported seeing adult male loggerheads during the late 1970s at Sekania, Zakynthos, Greece; White (2007) studied adult male loggerheads foraging at Kefalonia, Greece; and Schofield et al. (2010) reported that some male loggerheads are resident at Zakynthos National Marine Park, Greece, the largest confirmed loggerhead rookery in the Mediterranean. In Drini Bay, this unusual aggregation of males heightens its conservation importance as a critical habitat, especially if the threatened impact of global climate change is considered, which may force embryonic turtle sex ratios toward female dominance (Davenport 1989; Poloczanska et al. 2009).

Adult males are easily recognizable from the well-developed tail, which extends beyond the carapace's posterior margin. Casale et al. (2005), using similar methodology on male loggerheads captured in Italian waters, concluded that only turtles with CCL > 65.0 cm needed to be assessed. In contrast, we noted at Drini Bay that caudal development occurred in males at smaller size classes than those reported by Casale et al. (2005), so tail data were recorded from all captured sea turtles (White et al. 2008, 2010a, 2010b, 2011a, 2011b).

The development of secondary sexual characteristics in juvenile turtles occurs across a range of year classes, perhaps spanning a decade or more (Limpus and Limpus 2003; White 2007; White et al. 2010b, 2011a). Morphological changes that show the onset of maturation in males (proximal thickening and distal elongation of the tail) were observed in 76 loggerheads at Drini Bay. These animals were mostly in CCL size classes of 60 and 70 cm; the smallest turtle showing clear tail development had a CCL of 52.0 cm.

Regarding the reported overlap of 12.5 cm in the CCL size range between the smallest adult and the largest subadult, it is unclear how many years of growth this overlap represents. Research by van Dam in Puerto Rico found considerable size dimorphism in adult male hawksbill turtles; 1 particular photograph shows 2 males, clearly adult with very long tails, but 1 has a CCL of 52.0 cm, the other a CCL of 83.0 cm (van Dam and Diez 1998).

Short-Tailed Turtles

Initially, it was difficult to know whether adult females used Drini Bay; there are no confirmed nesting observations for Albania or the southeastern Adriatic Sea (Groombridge 1990; Casale and Margaritoulis 2010; Haxhiu 2010; White et al. 2010a, 2011a, 2011b). Several short-tailed loggerheads (n  =  53) were adult-sized animals, and their dates of capture coincided with the known egg-laying period in the Ionian Sea, mostly June–August (Groombridge 1990; Marquez 1990; Margaritoulis et al. 2003; White 2007; Casale and Margaritoulis 2010). Assuming that males and females do actually mature at similar sizes (Dodd 1988; Limpus and Limpus 2003; Casale et al. 2005), the CCL size classes of adult males captured in Drini Bay were used as a reference point for female maturity; there was no possibility of using laparoscopy. Loggerheads with CCL ≥ 70 cm can reasonably be assumed to be adults, even if some turtles in the Mediterranean mature at slightly smaller sizes than this. Margaritoulis et al. (2003) gave the average minimum size of nesting females in Cyprus, Greece, Libya, Tunisia, and Turkey as CCL  =  69.0 cm (mean CCL  =  78.0 cm). Broderick et al. (2003) reported that their smallest egg-laying female in Cyprus had a CCL of 63.0 cm (mean CCL  =  73.6 cm).

Further evidence for size at maturity was collected from Drini Bay when a loggerhead (CCL  =  70.0 cm) aborted eggs in the water while detained overnight (July 2010). Another loggerhead (CCL > 70 cm) was captured in stavnikes in 2 consecutive years (26 June 2008 and 3 July 2009), with “mating scars” on the back of the neck each time, suggesting that she could have been undertaking reproductive migrations.

A similar approach to that outlined above for adults was used to categorize sexually developing females, although this was more difficult, as development is mostly internal. Subadult males (i.e., turtles showing definite tail development) ranged in size from 52.0 to 81.0 cm CCL (mean CCL  =  67.5 cm; SD  =  ± 5.8; n  =  76). Selecting a threshold of CCL > 60 cm, a further 108 turtles may therefore be sexually maturing females; clearer evidence will be obtained from any subsequent recaptures. The tail tips of all turtles in the 2 smallest CCL size classes (30 and 40 cm) were subcarapace; that is, there were no obvious external signs of sexual development, so these could be juveniles of either sex.

This method is very simple to use in the field when laparoscopy is unavailable and needs only a tape measure. We found 1 shortcoming that concerned the third tail measurement in a small number of turtles. The tail tips of 2 apparent female turtles (70-cm size class) had extended beyond the carapace margin by 0.5 and 1.0 cm, respectively, but they could not be described as “long-tailed” animals. Similarly, for 5 clearly male loggerheads, the tail tip had not yet reached the rearmost carapace margin; in 1 animal, the distance between the plastron and carapace rear margins was very long and thus not a good guide; another had a very distorted carapace—the supracaudal scutes were displaced about 15.0 cm to the right, and accurate measurements were not possible. Finally, several turtles had fractured or missing marginal scutes and thus lacked important landmarks for measuring.

Incremental Growth

Polyphasic growth (i.e., growth spurts) has been discussed by Chaloupka (1998). Turtle growth rates reported here show that the tails grew considerably faster than the carapaces. In other words, there were differential growth rates between the developing secondary sexual characteristics compared with increases in carapace length and width. Unfortunately, there were only 4 records for tail growth, compared with 26 for carapacial growth. A more accurate size for the onset of sexual development and subsequent maturity could be determined by widening the sample size for caudal incremental growth to include different species and populations or by confirming sex using laparoscopy.

Marine Habitat Use

The mean CCL for 176 loggerheads measured at Centro Recupero Tartarughe Marine Lampedusa (White 2007) was at least one 10-cm size class smaller than those reported here for Drini Bay, suggesting that Lampedusan sea areas act as an oceanic habitat (even though water depth there is shallower than 200 m), whereas northern Albania is a neritic benthic habitat (Bolten 2003). Furthermore, there were no animals larger than the 80-cm size class from either Lampedusa or Drini Bay, supporting turtles in the Mediterranean being of smaller size than those from other parts of the world (Marquez 1990; Limpus et al. 1994, 2001; Margaritoulis et al. 2003).

Casale et al. (2007) concluded that loggerheads in the Mediterranean shift from an oceanic habitat to a neritic phase at some particular place that is then used for foraging throughout their life, forsaking other neritic areas. The Drini Bay study does not support that idea for 2 main reasons: the number of turtles recaptured interannually was very small (n  =  32 recaptures from 2008 to 2010), and most of the turtles in the bycatch are untagged, suggesting that there is a continuous exchange of foraging animals using the bay. White (2007) found a similar pattern during year-round research (2000–2004) in foraging areas at Kefalonia, Greece, with only 1 loggerhead being encountered there in more than 1 yr. This suggests that loggerheads remain opportunistic foragers, moving from place to place to maximize food availability, but that they will remain with a food source for some weeks or even months before emigrating elsewhere (Polovina et al. 2000, 2004, 2006; White 2007; White et al. 2010a, 2011a).

Twenty-five turtles were recaptured at Drini Bay within the same field season (n₂₀₀₈  =  16; n₂₀₀₉  =  9), suggesting that these individuals showed short-term site fidelity by remaining in local waters for several weeks. White (2007) had also reported similar behavior from Kefalonia, where some loggerheads remained there, foraging for up to 26 wk before moving elsewhere. In terms of incremental growth, within-season recaptures were not expected to have grown noticeably in this time.

Population Structure

Drini Bay appears to be particularly important as a developmental habitat for subadult males (n  =  76), subadult females (n  =  108), late-stage short-tailed juveniles (n  =  78), and smaller turtles (n  =  20). In total, 282 turtles (69% of captured turtles) had apparently not yet reached maturity.

Adults also make up a substantial part of the population: 35 adult males and 53 probably adult females. The most likely reason for turtles to be in Drini Bay is foraging given the absence of nesting reports from the Adriatic Sea (Casale and Margaritoulis 2010). The shallow, sandy bay is rich in benthic fauna; however, as the underwater visibility is 0, we have no direct observations of turtles feeding. Quietness and low incidence of people is not usually mentioned as a habitat attribute, but both of these are the case in the southern part of Drini Bay and at Patoku, and we suggest that these may make an important contribution to the presence of turtles here.

The congregation of adult males and females means that mating is a possibility, whether in Drini Bay or out in the Adriatic Sea. White et al. (2010b, 2011a, 2011b) noted that this was an unusual assemblage of adult males, as the nearest known nesting sites are farther south in the Ionian Sea. This suggests that females would have to travel northward to copulate before migrating to their nesting beaches; however, there may be as-yet-undiscovered nesting locations in Albania or in the Adriatic region.

An interesting finding is that all of the previously tagged loggerheads re-encountered here (n  =  32) had originally been tagged in Drini Bay; no turtles possessed flipper tags applied elsewhere. In contrast, some loggerheads in Croatia had been tagged while nesting at Zakynthos, Greece (Margaritoulis et al. 2003; Lazar et al. 2004b), and these would have had to migrate past or through Drini Bay on their way north (Lazar et al. 2003). So does this mean that loggerheads really are “opportunistic” for foraging, for habitat selection, and for their migratory routes? Lohman et al. (1997) suggested that there was no common migratory pathway for C. caretta and that individuals migrated from diverse starting points only to converge at various mating or nesting habitats. We would also suggest convergence at rich foraging areas, such as Drini Bay.

Only 3 turtles tagged in Drini Bay have been reported from elsewhere. A diver saw a loggerhead turtle in the Gulf of Sirte, Libya (November 2010; tag no.AL0045; 16/06/2009). ARCHELON provided the first confirmation (21 June 2011) that a tagged loggerhead had nested at Sekania Beach, Zakynthos National Marine Park, Greece (tag no. AL0127; 31/07/2009). The reported CCL shows that incremental growth was 3.5 cm in nearly 2 yrs. A third loggerhead also nested at Zakynthos in July 2012 (Archelon, unpubl. data); its exact tag number was misread (number inverted). This suggests that a novel life history link between these 2 sites may have been discovered; a recent genetic study also supports this conclusion (Yilmaz et al. 2012).

The population structure suggested here for Drini Bay should be treated cautiously, as most turtles were taken as bycatch from just 2 stavnikes during the summer months (2008–2010). There are other stavnikes in the 30-km bay (White et al. 2011a, 2011b) as well as other fishing activities (e.g., trawling and gill nets). The number of turtles that utilize the bay is probably substantially higher, as it seems unlikely that every turtle foraging locally would enter 1 of these 2 traps.

Green Turtles

The presence of live green turtles in each year was of special importance. Lazar et al. (2004b) reported only dead animals and historical records, and those reported from Albania were mostly thought to be smaller than 50.0 cm CCL (Haxhiu 2010), although it appears now that some of these may have been misidentified juvenile C. caretta (White et al. 2011a). Regionally, C. mydas nests only in the northeastern Mediterranean (Cyprus, Turkey, and Levant) and more usually has a tropical distribution (White et al. 2011a). The present authors added Albania as a CMS (Bonn Convention 1979) range state for C. mydas in the Mediterranean (White et al. 2011a). Bycatch in 2010 included 2 larger C. mydas that appeared to be subadult males; these have been reported as a new life stage for this species from the Adriatic Sea (White et al. 2011b).