Skip to main content
NCBI home page
Journal of Anatomy logoLink to Journal of Anatomy
. 2019 Jun 14;235(4):836–846. doi: 10.1111/joa.13022

Supernumerary scutes verify a segment‐dependent model of the horny shell development in turtles

Gennady Cherepanov 1,✉, Yegor Malashichev 1, Igor Danilov 2
PMCID: PMC6742898  PMID: 31198986

Abstract

Turtle horny shell has a scute pattern, which is conservative through evolution and across species. The discovery of epidermal placodes as the scute primordia and their strict topographical association to the somites of the turtle embryo suggested a new interpretation of the developmental mechanism of the scute pattern. Here, we tested the hypothesis that horny scutes develop from a mosaic of placodes corresponding exactly to the paths of myoseptae, with vertebral and pleural scutes developing staggered in adjacent segments, and marginal scutes developing in every segment. This scheme predicts little variation in marginals and suggests intercalary supernumerary scutes as potential variations for the vertebral and pleural rows. We examined spatial and numerical variations of the horny shell in 655 newly hatched olive ridley sea turtle, Lepidochelys olivacea, which is known to have a highly variable horny shell. In total, 120 patterns of carapacial scutes and 10 patterns of scutes on plastron, differing in the number and position of scutes were found. The number of vertebral scutes varied from 4 to 10. Variations with five, six and seven vertebrals occurred with the greatest and nearly equally frequency (31.5% on average). Pleural scutes were from 5 to 10 at one or both sides, and the typical symmetric pattern for sea turtles with five pairs of pleurals was only seen in ca. 12% of specimens. In contrast, the majority of the specimens (92.7%) had just 13 pairs of marginals, showing a stable normal pattern. Similarly, on plastron the horny scutes were conservative, too; about 85% of specimens standardly had six pairs of plastral scutes and all specimens had four pairs of inframarginals. Despite a high level of variation of vertebral and pleural scutes in olive ridley turtle, all patterns fall into the theoretical spectrum of possible variants predicted by the segment‐dependent model of development of the turtle horny shell. Therefore, the results of our analysis support the existence of direct morphogenetic correlation between the number and distribution of normal and supernumerary scutes and metamere organization of the turtle embryo.

Keywords: Lepidochelys olivacea, scute variation, segment‐dependent model, turtle shell

Introduction

Turtles have a unique pattern of the horny scutes (pholidosis) of the shell (Fig. 1). This pattern demonstrates a rather high evolutionary stability and at the same time a wide range of individual variability (Zangerl, 1969). Numerous specimens of various turtle species with anomalies of the scutes have been reported in the literature (Gadow, 1899; Parker, 1901; Newman, 1906; Coker, 1910; Deraniyagala, 1939; Lynn, 1937; Zangerl & Johnson, 1957; Douglass, 1977; Ewert, 1979; Mast & Carr, 1989; Fernandez & Rivera, 2004; Özdemir & Türkozan, 2006; Cordero‐Rivera et al. 2008; Davy & Murphy, 2009; Velo‐Antón et al. 2011; Cherepanov, 2014; and others). It is generally believed that such variability is caused by negative environmental effects: extreme temperatures or humidity, as well as pollution may affect embryonic development, resulting in different kind of anomalies (Lynn & Ullrich, 1950; Bujes & Verrastro, 2007; Velo‐Antón et al. 2011; Telemeco et al. 2013; Caracappa et al. 2016; Loehr, 2016; Zimm et al. 2017). However, some types of scute anomalies have a specific pattern on a species or population level, suggesting a genetic control (Zangerl & Johnson, 1957; Zangerl, 1969; Cordero‐Rivera et al. 2008; Velo‐Antón et al. 2011). Several studies have shown that shell scute anomalies in turtles often are correlated with each other; thus has allowed some authors to put forward a number of assumptions regarding the morphogenetic causes of the observed regularities (Gadow, 1899; Coker, 1910; Zangerl & Johnson, 1957; Ewert, 1979). However, these assumptions were rather speculative without a proper empirical support, since the development of the scutes was poorly known at that time (see Cherepanov, 2015, for details). Special studies of the embryonic development of the turtle horny shell have been provided only in the past few decades (see Moustakas‐Verho & Cherepanov, 2015). The discovery of epidermal placodes as the scute primordia (Cherepanov, 1989) and description of their development in some species of turtles shed light on fundamental mechanisms of the formation of their scute pattern. In particular, a strict topographical association of the scute primordial placodes and somites of the turtle embryo was found. Based on obtained morphogenetic data, the segment‐dependent model of development of the turtle horny shell (hereinafter the segment‐dependent model), was established, explaining both ontogenetic and evolutionary stability of its general arrangement, and revealing morphogenetic sources of the scute anomalies (Cherepanov, 1989, 2002, 2005, 2006, 2014, 2015).

Figure 1.

Figure 1

Open in a new tab

The scheme of the general pattern of the horny shell in sea turtles (Cheloniidae) and nomenclature of scutes. (A) Carapace scutes: Ce, cervical; M, marginals; P, pleurals; V, vertebrals. (B) Plastron scutes: Ab, abdominals; An, anals; Ax, axillaries; Fe, femorals; Gu, gulars, Hu, humerals; Im, inframarginals; Pe, pectorals; 1—13, serial numbers of scutes.

The main provisions of the segment‐dependent model are as follows:

  1. Shell scutes arise during turtle embryogenesis as small, locally placed epidermal placodes (Fig. 2A). The placodes represent morphogenetic modules, which develop relatively autonomously, whereas the scute pattern represents a module system, which was confirmed on the genetic level (see Moustakas‐Verho et al. 2014).

    Figure 2.

    Figure 2

    Open in a new tab
    (A) Position of the scute epidermal placodes in carapace and plastron of the turtle embryo. (B) Frontal section of the embryonic carapace of Testudo graeca (stage 14 after Yntema, 1968), illustrating pleural placodes development in the bottom of septal invaginations. You can see a vacant septal invagination between two others, which are occupied by placodes. (C) Transverse section of the embryonic carapace of Testudo graeca (stage 15) in the region of the pleural and marginal placodes. (D) Frontal section of the embryonic carapace of Emys orbicularis (stage 14), illustrating position of the marginal placodes in septal invaginations of the carapacial ridge, opposite the ribs. Ab.P, abdominal placode; An.P, anal placode; Ce.P, cervical placode; C.R, carapace ridge; Fe.P, femoral placode; Gu.P, gular placode; Hu.P, humeral placode; M.P, marginal placode; P.P, pleural placode; Pe.P, pectoral placode; R, rib; V.P, vertebral placode; 1—12, serial numbers of placodes. Sections are stained with Delafield's hematoxylin with eosin. Scale bar: 250 μm.

  2. In the carapace, the placodes arise exclusively on the bottom of the intersegmental invaginations of the embryo (hereinafter septal invaginations), located in the area of the transverse trunk myosepta (=ribs; Fig. 2B). Therefore, the position of the placodes is determined by a primary somitic segmentation of the embryo.

  3. The marginal placodes are developed in each septal invagination (Fig. 2C,D) and their number (12–13 pairs) corresponds to the number of myosepta, separating myomeres of the carapace (herein, the numeration of the carapacial myosepta corresponds to the numeration of the marginal scutes). Five vertebral placodes and four to five pairs of the pleural placodes are located with a pass through one septal invagination in the areas of even and odd myosepta, respectively, i.e. staggered relative to each other (Fig. 2A).

  4. The cervical and vertebral scutes arise as paired (left and right) primordia, normally fusing with each other.

  5. The vertebral, pleural and, probably, plastral rows of scutes are characterized by presence of the vacant (normally free from the scute primordia of the same row) septal invaginations (myosepta), which suggests a reason for the asymmetric pattern and appearance of the supernumerary (intercalary) scutes in these rows.

  6. The number of scutes in the longitudinal row cannot be more than one per a trunk segment (= myoseptum) and the total number of scutes in the longitudinal row cannot exceed the number of the trunk segments, along which this row is formed.

The present study aims to verify the main provisions of the segment‐dependent model using original data on scute variability. As the object of study, the olive ridley sea turtle (Lepidochelys olivacea) has been chosen, due to its high natural level of the scute variation and ‘truly polymorphic carapace’ (Pritchard, 1979). Unlike all other turtle species with a single dominating (normal) scute pattern and rare anomalies, L. olivacea does not have a single normal scute pattern because the frequency of anomalies is extremely high and the range of variation is wide. For these reasons, L. olivacea is considered the best model, which is able to provide a mass material on scute anomalies without negative experimental effect, for example, higher incubation temperature (see Zimm et al. 2017). Such an effect is known strongly to disrupt the normal course of the embryogenesis, affecting, among other things, somitic segmentation (Primmett et al. 1989), which is not acceptable for the purposes of our study.

Materials and methods

The carapace and plastron scutes of 655 L. olivacea hatchlings from nine natural clutches were examined. The study was conducted during two seasons (November–February 2015 and November–December 2016) on the Southern coast of Sri Lanka near the town of Kosgoda. Turtle eggs were obtained from natural nesting sites, and transferred to the territory of the turtle farm in Kosgoda. The incubation site was located on a sandy beach 50 m from the water's edge. The eggs were incubated in sand 30–50 cm deep (30–40 eggs per nest with indication of the natural clutch number), at ambient temperature. Living newborn turtles were photographed in dorsal and ventral aspect and then released into nature. Photographing was performed using a Canon EOS 500D camera and Sigma AF 18–250 mm F3.5–6.3 HSM camera lens to get high resolution images. The photographs were viewed under high magnification to obtain a detailed description of the scute pattern. The data obtained were summarized in a table and analyzed. The statistical analysis of the available data on different nests and estimation of the female contribution to the pattern of the scute variation will be published in a separate paper devoted to structure of the examined population of L. olivacea.

The nomenclature of the scutes is presented in Fig. 1. The scutes of the main part of the plastron are usually represented by six pairs (gulars, humerals, pectorals, abdominals, femorals, anals), referred to as the plastral scutes, or plastrals. The numerous axillary scutes characteristic of sea turtles only are not considered in our study.

Results

In total, among L. olivacea specimens, 120 scute patterns, distinguished by the number of carapacial scutes, were found (Table 1, Figs 3A and 5A–C). Symmetrical patterns are represented by 34 variants in 399 specimens (60.9%) with a frequency of up to 11.9% (Table 1, pattern 2). The number of asymmetric patterns is higher, 86 variants, but their individual frequency does not exceed 4.1% (Table 1, pattern 27).

Table 1.

Scute patterns of the carapace and their frequency (%) in 655 newborn Lepidochelys olivacea

No Ce‐LM‐LP‐V‐RP‐RM n (%)
1 1‐12‐6‐4‐6‐12 1 (0.15)
2 1‐13‐5‐5‐5‐13 78 (11.9)
3 1‐10‐5‐5‐6‐12 1 (0.15)
4 1‐13‐5‐5‐6‐13 5 (0.75)
5 1‐13‐6‐5‐5‐13 5 (0.75)
6 0‐13‐6‐5‐6‐13 2 (0.3)
7 1‐13‐6‐5‐6‐13 47 (7.2)
8 1‐14‐6‐5‐6‐13 1 (0.15)
9 2‐13‐7‐5‐5‐13 1 (0.15)
10 1‐13‐6‐5‐7‐13 6 (0.9)
11 1‐13‐7‐5‐6‐13 15 (2.3)
12 1‐13‐6‐5‐8‐13 1 (0.15)
13 0‐13‐7‐5‐7‐13 1 (0.15)
14 1‐13‐7‐5‐7‐13 22 (3.4)
15 1‐13‐7‐5‐8‐13 5 (0.75)
16 1‐13‐8‐5‐7‐13 2 (0.3)
17 1‐13‐8‐5‐8‐13 6 (0.9)
18 1‐13‐5‐6‐5‐13 1 (0.15)
19 1‐13‐5‐6‐6‐13 7 (1.0)
20 1‐13‐6‐6‐5‐13 2 (0.3)
21 1‐13‐7‐6‐5‐13 1 (0.15)
22 1‐13‐6‐6‐6‐12 1 (0.15)
23 1‐13‐6‐6‐6‐13 56 (8.5)
24 2‐13‐6‐6‐6‐13 1 (0.15)
25 1‐14‐6‐6‐6‐13 2 (0.3)
26 0‐13‐6‐6‐7‐13 1 (0.15)
27 1‐13‐6‐6‐7‐13 27 (4.1)
28 1‐14‐6‐6‐7‐13 1 (0.15)
29 0‐13‐7‐6‐6‐13 4 (0.6)
30 1‐13‐7‐6‐6‐13 24 (3.7)
31 1‐14‐7‐6‐6‐14 1 (0.15)
32 0‐13‐7‐6‐7‐13 4 (0.6)
33 1‐13‐7‐6‐7‐13 23 (3.5)
34 2‐13‐7‐6‐7‐13 2 (0.3)
35 1‐13‐7‐6‐7‐14 1 (0.15)
36 0‐13‐6‐6‐8‐13 2 (0.3)
37 0‐13‐8‐6‐6‐13 1 (0.15)
38 1‐13‐7‐6‐8‐13 4 (0.6)
39 1‐13‐8‐6‐7‐13 3 (0.45)
40 2‐14‐8‐6‐7‐13 1 (0.15)
41 1‐13‐8‐6‐8‐13 9 (1.35)
42 1‐13‐8‐6‐9‐13 1 (0.15)
43 1‐14‐8‐6‐9‐13 1 (0.15)
44 1‐13‐9‐6‐7‐13 2 (0.3)
45 1‐13‐9‐6‐8‐13 1 (0.15)
46 1‐13‐9‐6‐9‐13 1 (0.15)
47 2‐13‐9‐6‐9‐13 1 (0.15)
48 1‐13‐5‐7‐6‐13 1 (0.15)
49 1‐13‐5‐7‐6‐14 1 (0.15)
50 1‐13‐6‐7‐5‐13 1 (0.15)
51 1‐13‐5‐7‐7‐13 1 (0.15)
52 1‐14‐7‐7‐5‐14 1 (0.15)
53 0‐13‐6‐7‐6‐13 1 (0.15)
54 1‐13‐6‐7‐6‐13 23 (3.5)
55 2‐13‐6‐7‐6‐13 1 (0.15)
56 1‐13‐6‐7‐6‐14 1 (0.15)
57 1‐14‐6‐7‐6‐13 1 (0.15)
58 1‐13‐6‐7‐7‐13 23 (3.5)
59 2‐13‐6‐7‐7‐13 3 (0.45)
60 1‐14‐6‐7‐7‐14 1 (0.15)
61 1‐13‐7‐7‐6‐13 25 (3.8)
62 2‐13‐7‐7‐6‐13 2 (0.3)
63 1‐13‐7‐7‐6‐14 1 (0.15)
64 1‐14‐7‐7‐6‐14 1 (0.15)
65 1‐13‐6‐7‐8‐13 2 (0.3)
66 1‐13‐8‐7‐6‐13 1 (0.15)
67 2‐13‐8‐7‐6‐13 1 (0.15)
68 2‐14‐8‐7‐6‐14 1 (0.15)
69 0‐13‐7‐7‐7‐13 3 (0.45)
70 0‐14‐7‐7‐7‐13 1 (0.15)
71 1‐12‐7‐7‐7‐12 1 (0.15)
72 1‐12‐7‐7‐7‐13 1 (0.15)
73 1‐13‐7‐7‐7‐12 1 (0.15)
74 1‐13‐7‐7‐7‐13 68 (10.4)
75 2‐13‐7‐7‐7‐13 3 (0.45)
76 1‐13‐7‐7‐7‐14 1 (0.15)
77 1‐14‐7‐7‐7‐13 2 (0.3)
78 0‐13‐7‐7‐8‐13 1 (0.15)
79 1‐13‐7‐7‐8‐13 10 (1.5)
80 1‐14‐7‐7‐8‐14 1 (0.15)
81 0‐13‐8‐7‐7‐13 2 (0.3)
82 1‐13‐8‐7‐7‐13 6 (0.9)
83 2‐13‐8‐7‐7‐13 2 (0.3)
84 1‐13‐8‐7‐7‐14 1 (0.15)
85 1‐14‐8‐7‐7‐14 3 (0.45)
86 1‐13‐7‐7‐9‐13 1 (0.15)
87 1‐13‐8‐7‐8‐12 1 (0.15)
88 1‐13‐8‐7‐8‐13 15 (2.3)
89 2‐13‐8‐7‐8‐13 2 (0.3)
90 1‐13‐8‐7‐8‐14 1 (0.15)
91 1‐14‐8‐7‐8‐13 2 (0.3)
92 1‐14‐8‐7‐8‐14 1 (0.15)
93 1‐13‐8‐7‐9‐13 3 (0.45)
94 1‐13‐8‐7‐9‐14 1 (0.15)
95 2‐14‐8‐7‐9‐14 1 (0.15)
96 1‐13‐9‐7‐8‐13 2 (0.3)
97 1‐14‐9‐7‐8‐14 1 (0.15)
98 1‐13‐9‐7‐9‐13 3 (0.45)
99 1‐13‐6‐8‐5‐13 1 (0.15)
100 2‐13‐6‐8‐5‐13 1 (0.15)
101 1‐15‐6‐8‐7‐14 1 (0.15)
102 1‐13‐7‐8‐6‐13 2 (0.3)
103 2‐13‐7‐8‐6‐13 1 (0.15)
104 1‐14‐7‐8‐6‐14 2 (0.3)
105 1‐13‐7‐8‐7‐13 2 (0.3)
106 2‐13‐7‐8‐7‐13 1 (0.15)
107 1‐14‐7‐8‐7‐14 1 (0.15)
108 1‐13‐7‐8‐8‐13 2 (0.3)
109 2‐13‐7‐8‐8‐13 1 (0.15)
110 1‐13‐7‐8‐8‐14 1 (0.15)
111 1‐13‐8‐8‐7‐13 3 (0.45)
112 2‐13‐8‐8‐7‐13 2 (0.3)
113 2‐14‐8‐8‐7‐13 1 (0.15)
114 1‐13‐8‐8‐8‐13 7 (1.0)
115 1‐14‐8‐8‐8‐14 1 (0.15)
116 2‐14‐8‐8‐8‐14 1 (0.15)
117 1‐13‐9‐8‐9‐13 2 (0.3)
118 1‐14‐9‐9‐10‐14 1 (0.15)
119 1‐13‐9‐10‐9‐13 1 (0.15)
120 1‐13‐9‐10‐10‐13 1 (0.15)
Open in a new tab

Ce, cervicals; LM, left marginals; LP, left pleurals; RP, right pleurals; RM, right marginals; V, vertebrals.

Figure 3.

Figure 3

Open in a new tab

Some examples of carapace scute variation in newborn Lepidochelys olivacea. (A) Specimen (P2175395) has seven vertebral and nine pairs of pleural scutes. (B) Specimen (P2124568) shows incomplete fusion of the cervical and vertebral 1 scute. (C) Specimen (P2175495) has paired cervical scutes and an supernumerary scute between them and vertebral row. (D) Specimen (P2175658) shows position of the supernumerary pleural scute between the regular ones. (E) Specimen (P2205775) has supernumerary pair of marginals.

The cervical scutes are represented by unpaired (91.9%) oring paired (4.6%) variants (Fig. 3A,C). Some specimens (3.5%) lack cervicals, with their place occupied by an elongated vertebral scute, probably as a result of fusion of the cervical and vertebral 1 (Fig. 3B).

The number of vertebrals varies from 4 to 10. A series of five vertebrals, typical for most turtle species, was found in 30.2% of L. olivacea specimens. Variants with six and seven vertebrals have a similar frequency (28.4 and 35.8%). Further increase in number of vertebrals is rarer. The maximal number of 10 vertebrals is present only in two specimens (Table 1, patterns 119–120). Supernumerary vertebrals, extra between regular scutes of the row, appear in the caudal part of the carapace more often (60%) than in the cranial part. Conditionally, a rudimentary supernumerary scute between the cervical and vertebral 1 can be also assigned to the vertebral row (Fig. 3C).

The number of pleural scutes in a row varies from 5 to 10. Five pairs of pleurals, typical for some sea turtles, were found only in 11.9% of specimens, whereas six and seven pairs of pleurals each were observed in 21.1% of specimens, and eight and nine pairs in 7 and 1/2% of specimens, respectively. The rest of the specimens have various asymmetric variants of pleurals. Like vertebrals, supernumerary pleurals appear more frequently in the caudal part of the carapace than in the cranial part (more than 85% of specimens have supernumerary pleural between regular pleurals 4 and 5). Usually, the supernumerary scutes in the caudal part of the carapace are the same size as the regular ones, whereas the supernumerary scutes in the cranial part frequently look like small wedge‐shaped incuts between the regular scutes (Fig. 3D). Despite high diversity of the scute patterns, patterns with symmetric pleurals prevail over asymmetric ones in L. olivacea: the same number of pleurals on the left and right sides was found in 62.4% of specimens.

The number of marginal scutes varies from 10 to 15. Most of specimens (92.7%) have 13 pairs of marginal, which can be considered the norm. A reduced number of marginals was observed in 1.1% of specimens. A more frequent deviation from the norm is the presence of one or a pair of supernumerary marginals (Fig. 3E) in the caudal (5.6%) or cranial (0.6%) parts of the carapace. In most cases, supernumerary marginals are found in the specimens with supernumerary pleural scutes (seven or more).

Only 10 plastral scute patterns were revealed in L. olivacea specimens (Table 2). Most specimens (85.6%) have six regular pairs of plastrals (gulars, humerals, pectorals, abdominals, femorals and anals), i.e. a stable norm (Fig. 4A). Supernumerary elements are represented in most cases by paired or unpaired intergular scutes (Fig. 4B,C), found in 11.6% of the specimens. A rarer variant (1.8%) is the presence of a single supernumerary scute or a pair of them at the posterior edge of the plastron (Fig. 4D). The presence of additional scutes between the regular plastrals is even rarer. Such paired and unpaired intercalary scutes were found only in eight specimens (1.2%) of L. olivacea and only in the area between the humerals and pectorals (Fig. 4E).

Table 2.

Scute patterns of the plastron and their frequency (%) in 655 newborn Lepidochelys olivacea

No. LIM‐IG‐LPl‐RPl‐IAN‐RIM n (%)
1 4‐0‐6‐6‐0‐4 561 (85.6)
2 4‐1‐6‐6‐0‐4 42 (6.4)
3 4‐0‐6‐6‐1‐4 8 (1.2)
4 4‐1‐6‐6‐1‐4 1 (0.15)
5 4‐2‐6‐6‐0‐4 32 (4.9)
6 4‐0‐6‐6‐2‐4 2 (0.3)
7 4‐2‐6‐6‐1‐4 1 (0.15)
8 4‐0‐7‐6‐0‐4 2 (0.3)
9 4‐0‐6‐7‐0‐4 4 (0.6)
10 4‐0‐7‐7‐0‐4 2 (0.3)
Open in a new tab

IAN, interanals; IG, intergulars; LIM, left inframarginals; LPL, left plastrals; RIM, right inframarginals; RPL, right plastrals.

Figure 4.

Figure 4

Open in a new tab

Some examples of plastron scute variation in newborn Lepidochelys olivacea. (A) Specimen (P2124577) with a normal arrangement of scutes on its plastron. (B) Specimen (P2124537) has supernumerary unpaired intergular scute. (C) Specimen (P2205814) has supernumerary paired intergular scutes. (D) Specimen (P2175538) has supernumerary unpaired interanal scute. (E) Specimen (P2205831) shows the presence of a supernumerary scute between the humeral and pectoral scutes on the right side of its plastron.

The number of inframarginals, represented by four pairs, is very stable and was found in all examined specimens of L. olivacea.

Discussion

The key idea of the segment‐dependent model is a strict connection between the primary embryonic segmentation and position of scute primordia – epidermal placodes. As shown by embryologic studies (Cherepanov, 1989, 2002, 2005, 2006, 2014, 2015), the carapacial placodes arise exclusively in the area of the transverse trunk myosepta, which are marked by septal invaginations during early stages of the embryogenesis, and later by ribs (Fig. 2). Since the marginal placodes arise in each septal invagination, it can be concluded that the number of pairs of marginal scutes must correspond, as a rule, to the number of myosepta separating the carapacial segments. The idea of the quantitative correspondence between marginal scutes and carapacial metameres is not new. In the opinion of several authors (Gadow, 1899; Parker, 1901; Lynn, 1937), variability in the number of scutes of the marginal row is connected with change in the number of trunk segments, participating in the shell formation (excluding cases of the atypical scute fusion). At that, the number of the carapacial segments may increase or decrease, usually for no more than one segment from the norm. Most extant turtle species normally have 12 pairs of marginals. The norm of L. olivacea is increased to 13 pairs of the marginals, suggesting an increased number of the shell segments in comparison with other turtles. Aberrations in the number of the marginals in this species are relatively rare (7.3%), comparable to those in conservative (in the number of scutes) sea turtle such as Chelonia mydas (Özdemir & Türkozan, 2006; Ergene et al. 2011), and almost three times lower than, for example, those in the speckled tortoise Homopus signatus (Loehr, 2016). The revealed decrease in the number of marginals in seven specimens of L. olivacea is caused by partial (as can be easily diagnosed by rudiments of sulci) or full fusion of adjacent elements. Therefore, these variations are secondary and not connected with disorders in the segment‐dependent placode development, but they are consistent with the reaction‐diffusion model of Moustakas‐Verho et al. (2014). An increase in the number of marginals (14 or 15) was revealed in 41 specimens (6.3%). The increase is due to the addition of the supernumerary scutes on the cranial or caudal end of the marginal row (Fig. 3E). This is quite consistent with the idea of the possibility of the carapace elongation by inclusion into it of additional trunk segments (see above). Some turtle species, for instance fossil Boremys pulchra (Baenidae), normally have such extra scutes (Brinkman & Nicholls, 1991). In general, the available data on variability of marginals do not contradict the segment‐dependent model.

Usually, the turtle carapace has five vertebrals and four pairs of pleurals, which is significantly less in every row than the number of marginals. This is because the pleurals and vertebrals do not arise in ontogenesis in every septal invagination, but only in odd or even septal invaginations. As a result, these rows have vacant (free from scute placodes) septal zones, which can be ‘filled’ by supernumerary placodes as a result of even slight disorders of embryogenesis. As a rule, such supernumerary placodes are named accessory scutes between regular ones (Fernandez & Rivera, 2004; Cordero‐Rivera et al. 2008; Velo‐Antón et al. 2011; Saçdanaku & Haxhiu, 2016). Thus, the presence of the vacant septal zones explains the great variation in the number of pleurals and vertebrals. This is clearly observed in the specimens of L. olivacea under study (Fig. 5A,C,E). In this species, the variants with six and more scutes in the vertebral and pleural rows significantly prevail over the basic morphotype typical for most turtles (Table 1, pattern 2). Based on morphogenetic data, theoretically, the maximum number of pleurals and vertebrals in one row in L. olivacea may reach 11 if all potential septal zones are occupied by scute placodes (Fig. 6A). Patterns with 10 pleurals and vertebrals that are close to the possible maximum were recorded by us in three specimens of L. olivacea (Table 1, patterns 118–120); the sum of these internal carapacial scutes reaches 29. This value slightly exceeds those reported by Deraniyagala (1939, 26 scutes) but is within a theoretically possible range. The presence of 10 vertebrals and 10 pleurals in the norm was described only in the geoemydid genus Sakya known from the Neogene of Eastern Europe (Chkhikvadze, 1968). It is believed that such a multiple composition of the horny scutes is to some extent a recapitulation of the polymeric scute pattern in turtle ancestors (Newman, 1906; Grant, 1937; Cherepanov, 2006). It is important to note that the excess of the theoretically calculated maximum number of pleurals and vertebrals was not detected by us in the studied material or in publications of other authors.

Figure 5.

Figure 5

Open in a new tab

Newborn individuals of Lepidochelys olivacea with different arrangement of vertebral and pleural scutes in the carapace (A,B,C) and schemes of these structural variants with probable arrangement of their placodes in trunk segments (D,E,F). (A,D) Specimen (P2124572) with one supernumerary pleural scute on the right side of the carapace. (B,E) Specimen (P2175519) shows the symmetrical pattern with two supernumerary scutes in each pleural and vertebral row (C,F). Aberrant specimen (P2124575) with asymmetric number of the pleurals and supernumerary scute in the vertebral row. Supernumerary scutes are shaded; the probable position of placodes is shown by circles; the presumed position of the transverse myosepta is shown by a dotted line; 3—12, serial numbers of myosepta.

Figure 6.

Figure 6

Open in a new tab

The distribution of the supernumerary scutes on the carapace (A) and on the plastron (B) of Lepidochelys olivacea hatchlings and their frequency (%). Supernumerary scutes are shaded. No.‐No., distribution right – left.

As was shown by morphogenetic and genetic data (Cherepanov, 2006; Moustakas‐Verho et al. 2014), the vertebral placodes appear later than the pleural placodes and as paired (left and right) primordia. These primordia then fuse, forming five united elements. Signs of this fusion can be seen in the shell of adult turtles in a form of double centers of the scute growth (Grant, 1937). Thus, the developmental path of the scutes of the vertebral row confirms the possibilitity of its primary pairing (Gadow, 1899). Normally, the vertebral placodes arise symmetrically in areas of even trunk myosepta, i.e. in staggered order with the pleural placodes that are located in odd myosepta (Fig. 2A). As was shown by data on scute variability in many turtle species, the asymmetry of the pleurals often correlates with those of the vertebrals, which results in the formation of the anomalous pattern of the ‘zig‐zag’ type, or ‘dovetail’ syndrome (Zangerl & Johnson, 1957; Ewert, 1979; Pritchard, 2007; Davy & Murphy, 2009). It is worth mentioning that in such a case the staggered order of the pleurals and vertebrals on either side is retained more strictly than their bilateral symmetry (see Cherepanov, 1989, Fig. 5). The staggered order is disrupted only when the pleurals appear in each septal invagination in some part of the trunk. At that, the number of vertebrals in this area also increases. Such polymeric patterns are typical of the majority of the studied specimens of L. olivacea.

The plastral placodes in six pairs appear during ontogenesis later compared with the carapacial placodes, on the periphery of the plastron (Cherepanov, 2005, 2006). By the time the placodes appear, the plastron does not show clear signs of somitic segmentation, although the certain metameric organization in this area is observed in earlier stages of embryogenesis (Yntema, 1970; Guyot et al. 1994). If we assume that the mechanisms of determination of the position of the placodes in the plastron and carapace are identical, then, similarly to the carapacial placodes, the position of the plastral placodes must be connected with the somitic segmentation. Besides that, there are arguments in favor of the idea that plastral scutes arise not in each segment, but with a pass through one segment: (1) the majority of extant turtles normally have six pairs of plastral scutes, i.e. two pairs less than number of marginal pairs; (2) plastral scutes are approximately of the same size as the pleurals and vertebrals, either of which covers two adjacent segments; (3) adjacent to pectoral and abdominal scutes, there are four marginals or inframarginals, lying in each segment.

Results of the morphogenetic studies indicate that the horny parts of the plastron form independently from those of the carapace. This is also confirmed by numerous data on individual variability, including our own (Newman, 1906; Lynn, 1937; Zangerl & Johnson, 1957; Douglass, 1977; Cherepanov, 2014). First, the plastral scutes are less variable than the carapacial ones, and the frequency of their aberrations is significantly lower. In particular, L. olivacea was found to have only 10 different patterns of the plastral scutes (as opposed to 120 patterns of the carapacial scutes) and 85% of specimens show a stable norm (Table 2). Secondly, the carapacial anomalies, as a rule, are not accompanied by plastral ones. Thirdly, there is no correlation between scute variation of the carapace and those of the plastron.

Variability of the plastrals in L. olivacea is connected only with the appearance of the supernumerary elements (scute reduction was not observed). First of all, these are paired or unpaired supernumerary scutes located at the cranial and caudal end of the plastron (intergular and interanal scutes). It is interesting to note that some primitive turtles, as well as pleurodires, have intergular scutes as a norm (Zangerl, 1969). Another type of anomaly is the presence of the intercalary scutes inside the regular plastral row. Such symmetric or asymmetric scutes were found in eight L. olivacea specimens between the regular humeral and pectoral scutes (Fig. 4D,E). The origin of these scutes is likely the same as that of the supernumerary pleurals and vertebrals, i.e. connected with ‘filling’ of the vacant segmental zones. In general, the pattern of the plastrals is rather conservative. The lower degree of variation of the plastral scutes in comparison with the carapacial ones is probably explained by a simpler arrangement of the horny plastron and lower number of scutes in it (only six pairs of plastrals and four pairs of inframarginals).

As was shown above, the number and position of the marginals correspond to the number and position of the myosepta, separating segments of the carapace. This allows use of the marginal scute position for reconstruction of the primary segmentary composition of their shell. Determination of the position of the transverse trunk myosepta in the shell and their numeration (corresponding to numeration of the marginals) allow the segmentary position of other carapacial scutes and possibly plastral ones to be determined with a high probability. Besides that, connection of the scute position with certain trunk segments (in our case, the with number of myosepta) allow regular scutes, which have a normal position in relation to myosepta, to be distinguished from supernumerary scutes, which occupy areas of myosepta vacant as a norm (Fig. 5).

This method of identification of the scutes and their embryonic primordia positions in relation to transverse trunk myosepta was successfully used to explain the nature of the scute anomalies in the shell of Testudo graeca, a species with a well‐studied shell morphogenesis (Cherepanov, 2014). In the present paper, we use this method to compare the theoretically possible spectrum of scute variability (based on morphogenetic data) with real scute patterns found in specimens of L. olivacea (Table 1). This analysis demonstrated that the examined specimens of this species possess all theoretically possible positions of the carapacial scutes (Fig. 6A). At that, no specimen has a complete set of the regular and supernumerary carapacial scutes. As expected, scutes of the pleural and vertebral rows appear to be the most variable. In most cases, supernumerary scutes in these rows are represented by intercalary elements between the regular ones, and rarely by supernumerary elements on the periphery of the row. Even in the most polymeric pattern, the pleurals and vertebrals never exceed one scute per a segment (= myoseptum), and their number in the rows is always lower than the number of the marginal pairs (i.e. carapace segments).

Unlike the carapace, the plastron of L. olivacea implements only a small part of theoretically possible variants of the scute positions (Fig. 6B). The most common anomaly is the presence of supernumerary scutes on the cranial or caudal end of the plastron. The true intercalary scutes were found only in the area between the regular humeral and pectoral scutes (Fig. 4D,E).

However, in other turtle species, the intercalary scutes were described in other theoretically possible positions. For instance, in the Iberian population of Emys orbicularis, supernumerary scutes were found between the gular and humeral, humeral and pectoral, abdominal and femoral, and femoral and anal scutes (Cordero‐Rivera et al. 2008), i.e. in all possible positions. In addition, possibly similar anomalies were described in several species of the genus Graptemys (Zangerl, 1969: table 1). According to published data (e.g. Newman, 1906; Lynn, 1937; Deraniyagala, 1939; Zangerl & Johnson, 1957; Zangerl, 1969; Douglass, 1977; Cherepanov, 2005; Szczygielski et al. 2018), the maximal number of the plastrals in both fossil and living turtles never exceeds nine pairs, even as anomaly. This is corresponding completely to the segment‐dependent model discussed herein.

Despite the widest range of scute polymorphism in L. olivacea, all its carapacial and plastral scute patterns fit theoretically possible variants determined by the segment‐dependent model. Results of the study support existence of a strict morphogenetic correlation between the number and distribution of the shell scutes and metameric organization of the turtle embryo. The latter determines key features of the general pattern of the horny shell: the presence of no more than one scute per a trunk segment (= transverse myoseptum) in a longitudinal row and restriction of the scute number in a row by number of myosepta, separating segments of the shell. Thus, our study clearly demonstrates that the shell variability in the hatchlings of L. olivacea completely corresponds to the provisions of the segment‐dependent model.

Conclusions

We examined the variability of the horny shell in 655 newly hatched L. olivacea to verify the hypothesis that scutes develop from a mosaic of epidermal placodes exactly corresponding to the embryonic trunk segmentation (the segment‐dependent model of development of the turtle horny shell). We found 120 patterns of the carapacial scutes, distinguishing by number and distribution. Vertebral and pleural scutes showed the greatest variability due to the presence of vacant trunk segments in this region of the carapace, normally free from scute primordia. As a result, a wide range of variants of the supernumerary scutes was realized there. The scutes of the plastron are more conservative: only 10 patterns with supernumerary plastral scutes were found. Despite a high rate of the polymorphism of the scutes in L. olivacea, all their patterns fit theoretical prediction using the segment‐dependent model. The results of our study support a strict morphogenetic correspondence between the number and distribution of scutes and metamere organization of the turtle embryo.

Acknowledgements

The authors thank personally Mr. Dudley Perera, the Director of the Kosgoda Sea Turtle Conservation Project, and his staff for generous welcome. Special thanks are to volunteers E. E. Malashicheva (Russia) and J. Dechantsreiter (Germany) for help while conducting this study, and to E. M. Obraztsova (St. Petersburg, Russia) for some corrections and useful comments. The study was supported by Saint Petersburg State University (expedition grant No. 1.42.1095.2016) and Russian Foundation for Basic Researches (grant No. 18‐04‐01082).

References

  1. Brinkman DB, Nicholls EL (1991) Anatomy and relationships of the turtle Boremys pulchra (Testudines: Baenidae). J Vertebr Paleontol 11, 302–315. [Google Scholar]
  2. Bujes CS, Verrastro L (2007) Supernumerary epidermal shields and carapace variation in Orbigny's slider turtles, Trachemys dorbigni (Testudines, Emydidae). Rev Bras Zool 24, 666–672. [Google Scholar]
  3. Caracappa S, Pisciotta A, Persichetti MF, et al. (2016) Nonmodal scute patterns in the Loggerhead Sea Turtle (Caretta caretta): a possible epigenetic effect? Can J Zool 94, 379–383. [Google Scholar]
  4. Cherepanov GO (1989) New morphogenetic data on the turtle shell: discussion on the origin of the horny and bony parts. Stud Geol Salmanticensia Stud Palaeocheloniol 3, 9–24. [Google Scholar]
  5. Cherepanov GO (2002) Pholidosis of the turtle shell in onto‐ and phylogenesis. Zool Zh 81, 480–488. [Google Scholar]
  6. Cherepanov GO (2005) Turtle shell: Morphogenesis and Evolution. St Petersburg: St Peterb Gos Univ. [Google Scholar]
  7. Cherepanov GO (2006) Ontogenesis and evolution of horny parts of the turtle shell In: Fossil Turtle Research, Vol. 1 (eds Danilov IG, Parham JF.), pp. 19–33. St Petersburg: Russian Journal of Herpetology. [Google Scholar]
  8. Cherepanov GO (2014) Patterns of scute development in turtle shell: symmetry and asymmetry. Paleontol J 48, 1275–1283. [Google Scholar]
  9. Cherepanov GO (2015) Scute's polymorphism as a source of evolutionary development of the turtle shell. Paleontol J 49, 1635–1644. [Google Scholar]
  10. Chkhikvadze VM (1968) Sakyidae – a new family of fossil turtles. Paleontol J 2, 88–94. [Google Scholar]
  11. Coker R (1910) Diversity in the scutes of Chelonia. J Morphol 21, 1–75. [Google Scholar]
  12. Cordero‐Rivera A, Ayres Fernandez C, Velo‐Anto G (2008) High prevalence of accessory scutes and anomalies in Iberian populations of Emys orbicularis . Rev Esp Herpetol 22, 5–14. [Google Scholar]
  13. Davy CM, Murphy RW (2009) Explaining patterns of deformity in freshwater turtles using MacCulloch's hypothesis. Can J Zool 87, 433–439. [Google Scholar]
  14. Deraniyagala PEP (1939) Tetrapod Reptiles of Ceylon. Vol. 1. Testudinates and Crocodilians. London: Dubau & Co. [Google Scholar]
  15. Douglass JF (1977) Abnormalities of scutellation in a population of Gopherus polyphemus (Reptilia, Testudinaceae). Florida Sci 40(3), 256–257. [Google Scholar]
  16. Ergene S, Aymak C, Ucar AH (2011) Carapacial scute variation in green turtle (Chelonia mydas) and loggerhead turtle (Caretta caretta) hatchlings in Alata, Mersin, Turkey. Turk J Zool 35, 343–356. [Google Scholar]
  17. Ewert MA (1979) The embryo and its egg: development and natural history In: Turtles, Perspectives and Research (eds Harless M, Morlock H.), pp. 333–413. New York: John Wiley & Sons. [Google Scholar]
  18. Fernandez CA, Rivera AC (2004) Asymmetries and accessory scutes in Emys orbicularis from Northwest Spain. Biol Bratislava 59(Suppl 14), 85–88. [Google Scholar]
  19. Gadow H (1899) Orthogenetic variations in the shell of Chelonia In: Zoological Results based on Material from New Britain, New Guinea, Loyalty Islands and Elsewhere, Collected during the Years 1895, 1896, and 1897 (ed. Willey A.), part 3, pp. 207–222. Cambridge: University Press. [Google Scholar]
  20. Grant C (1937) Orthogenetic variation. Proc Indiana Acad Sci 46, 240–245. [Google Scholar]
  21. Guyot G, Pieau C, Renous S (1994) Developpment embryonnaire d'une tortue terrestre, la tortue d'Hermann, Testudo hermanni Gmelin, 1789. Ann Sci Nat Zool 15, 115–137. [Google Scholar]
  22. Loehr VJT (2016) Wide variation in carapacial scute patterns in a natural population of speckled tortoises, Homopus signatus . Afr J Herpetol 65, 47–54. [Google Scholar]
  23. Lynn WG (1937) Variation in scutes and plates in the boxturtle, Terrapene carolina . Am Nat 71, 421–426. [Google Scholar]
  24. Lynn WG, Ullrich M (1950) Experimental production of shell abnormalities in turtles. Copeia 253–262. [Google Scholar]
  25. Mast RB, Carr JL (1989) Carapacial scute variation in Kemp's Ridley sea turtle (Lepidochelys kempi) hatchlings and juveniles. Conserv Manag 202–219. [Google Scholar]
  26. Moustakas‐Verho JE, Zimm R, Cebra‐Thomas J, et al. (2014) The origin and loss of periodic patterning in the turtle shell. Development 141, 3033–3039. [DOI] [PubMed] [Google Scholar]
  27. Moustakas‐Verho JE, Cherepanov GO (2015) The integumental appendages of the turtle shell: an evo‐devo perspective. J Exp Zool (Mol Dev Evol) 324B, 221–229. [DOI] [PubMed] [Google Scholar]
  28. Newman HH (1906) The significance of scute and plate ‘abnormalities’ in Chelonia. Biol Bul 10, 68–114. [Google Scholar]
  29. Özdemir B, Türkozan O (2006) Carapacial scute variation in green turtle, Chelonia mydas hatchlings in Northern Cyprus. Turkish J Zool 30, 141–146. [Google Scholar]
  30. Parker GH (1901) Correlated abnormalities in the scutes any bony plates of the carapace of the sculptured tortoise. Am Nat 35, 17–24. [Google Scholar]
  31. Primmett DR, Norris WE, Carlson GJ, et al. (1989) Periodic segmental anomalies induced by heat shock in the chick embryo are associated with the cell cycle. Development 105, 119–130. [DOI] [PubMed] [Google Scholar]
  32. Pritchard PCH (1979) Encyclopedia of Turtles. Hong‐Kong: TFH Publ Inc. [Google Scholar]
  33. Pritchard PCH (2007) Evolution and structure of the turtle shell In: Biology of Turtles (eds Wyneken J, Godfrey MH, Bels V.), pp. 45–84. Boca Raton: CRC Press. [Google Scholar]
  34. Saçdanaku E, Haxhiu I (2016) Accessory scutes and asymmetries in European pond turtle, Emys orbicularis (Linnaeus, 1758) and Balkan terrapin, Mauremys rivulata (Valenciennes, 1833) from Vlora Bay, western Albania. Int J Fauna Biol Stud 3, 127–132. [Google Scholar]
  35. Szczygielski T, Słowiak J, Dróżdż D (2018) Shell variability in the stem turtles Proterochersis spp. PeerJ 6, e6134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Telemeco RS, Warner DA, Reida MK, et al. (2013) Extreme developmental temperatures result in morphological abnormalities in painted turtles (Chrysemys picta): a climate change perspective. Integr Zool 8, 197–208. [DOI] [PubMed] [Google Scholar]
  37. Velo‐Antón G, Becker CG, Cordero‐Rivera A (2011) Turtle carapace anomalies: the roles of genetic diversity and environment. PLoS ONE 6, e18714. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yntema CL (1968) A series of stages in the embryonic development of Chelydra serpentina . J Morphol 125, 219–251. [DOI] [PubMed] [Google Scholar]
  39. Yntema CL (1970) Extirpation experiments on embryonic rudiments of the carapace of Chelydra serpentina . J Morphol 132, 235–244. [DOI] [PubMed] [Google Scholar]
  40. Zangerl R (1969) The turtle shell In: Biology of the Reptilia. Vol. 1, Morphology (eds Gans C, Bellairs AD, Parsons TS.), pp. 311–339. London: Academic Press. [Google Scholar]
  41. Zangerl R, Johnson RG (1957) The nature of shield abnormalities in the turtle shell. Fieldiana Geol 10, 341–362. [Google Scholar]
  42. Zimm R, Bentley BP, Wyneken J, et al. (2017) Environmental causation of turtle scute anomalies in ovo and in silico. Integr Comp Biol 57(6), 1303–1311. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland

RESOURCES