Introduction
Research suggests that sexual dimorphism in the second-to-fourth digit ratio (2D:4D), when finger length is measured on the ventral side (i.e., the length between the fingertip and the proximal basal crease), is related with prenatal sex-hormone levels (1, 2). Individuals affected with conditions entailing high prenatal testosterone levels, such as congenital adrenal hyperplasia (3, 4), autism (5), or women with male co-twins (6), present lower 2D:4D ratios, measured ventrally, as compared with controls. Further, sex-chromosomal disorders, such as complete androgen insensitivity syndrome (7) and Klinefelter syndrome (8), support female-typical 2D:4D ratios in genetic males. Thus, the 2D:4D ratio, measured ventrally, might be related to prenatal sex-hormone levels. However, theoretically, digit ratios based on bone-length measurement, should be even more closely related to the prenatal sex-hormonal milieu, because it is the bone which determines most of the length of a finger, and sex hormones affect the posterior Hox genes, which are an important determinant of the relative growth of bones in general (9). This view is supported by studies reporting sexual dimorphism in the 2D:4D ratio based on bone length (10–13). However, probably because of the properties of non-hazardousness and easy measurement, digit ratios derived from ventral digit length have preferably been used in research.
Studies comprising several (i.e., alternative) digit ratios derived from ventral digit length have reported stronger sexual dimorphism occurring in digit ratios having digit 2 as one of the constituents of the ratio (14–17). On the other hand, studies comprising several digit ratios derived from metacarpal-bone length have reported stronger sexual dimorphism occurring in digit ratios having digit 5 as one of the constituents of the ratio [in gorillas (18); in humans (19)]. In addition, studies have also reported a pattern of similar sexual dimorphism in digit ratios (2D:4D) derived from metacarpal and phalangeal bone lengths (12). Thus, the pattern of sexual dimorphism in digit ratios derived from phalangeal bone length likely is different from the pattern of sexual dimorphism in the corresponding digit ratios, when these are derived from ventral digit-length measurements. Because sex differences in digit ratios from phalangeal bone length are noticeably stronger than those in the corresponding metacarpal-bone digit ratios [four times stronger with regard to the 2D:4D ratio (12)] and because the effect size for sex differences in metacarpal-bone digit ratios is close to d = 0.5 [for ratios with digit 5 as one of the digits (19)], a larger sexual dimorphism is expected for digit ratios from phalangeal bone lengths. Thus, dissimilar to prior studies which reported effects of small-to-medium-size [i.e., d = 0.2–0.5 (20)], the study of digit ratios derived from phalangeal bone length is expected to yield larger sex differences.
Prior related research has not attempted a study of the sexual dimorphism in all digit ratios derived from phalangeal bone length. One possible reason for this fact could be that common machines emit X-rays divergently from the source, and thus skewed images could appear for peripheral areas. Such X-ray images may contain systematic error when the variables of interest are the length ratios between length measures of focal and peripheral areas of the hand (for example, the ratio between digit 3 and digit 5 vs. the ratio between digit 3 and digit 4). Of note, McFadden and Bracht (18, 19) conducted both of their studies, involving the digit ratios of all metacarpal-bone lengths, on skeletons.
Similar to ventral finger-length measurement, dorsal finger-length measurement is easy, non-hazardous, and expected to be reliable (21). Dorsal digit length mainly includes fingertip fat and the length determined by the ray (i.e., chain) of the three (namely, proximal, intermediate, and distal) phalangeal bones. Studies have shown that size of fingertip fat is not sexually dimorphic and has no relationship with 2D:4D (10, 13, 22). As well, dorsal measurement of digit lengths does not include the finger flexion creases, which might be one factor of measurement error and unreliability. For these combined reasons, it is conceivable that, as compared to digit ratios derived from ventral digit length, digit ratios derived from dorsal digit length might more appropriately represent the actual sexual dimorphism effect in digit ratios derived from phalangeal bone length. In contrast to measuring phalangeal bone length, it is possible to measure accurately all dorsal digit lengths in a living hand. Hence, investigations of digit ratios derived from dorsal digit length may well be more suitable for a better understanding of the patterns of sexual dimorphism found in digit ratios.
Therefore, in the present series of studies, we measured dorsal digit length of all fingers (excluding the thumb) in both hands of study participants and compared the magnitude of the sexual dimorphism in the digit ratios calculated from these finger-length measurements. As elaborated on above, digit ratios derived from dorsal digit length may be suitable proxies of digit ratios derived from phalangeal bone length. The objective of the present research was therefore to understand the patterns of sexual dimorphism in digit ratios based on phalangeal bone length by studying dorsally measured digit ratios, which are expected to show patterns of sexual dimorphism similar to the one found in metacarpal-bone digit ratios (19). Accordingly, we hypothesized that stronger sexual dimorphism would occur in dorsally measured digit ratios involving digit 5 as one of the constituents of the ratio.
Studies have reported that sexual dimorphism in the 2D:4D ratio is established much earlier than puberty (2, 10, 23, 24). Similar to the case of the classic 2D:4D ratio, it is possible that sexual dimorphism in other (alternative) digit ratios also is determined early on. However, no prior research has focused on sexual dimorphism in alternative (other than 2D:4D) digit ratios among children [with the exception of the 3D:4D ratio (11)]. Moreover, studies also have shown that comparatively large age-related changes occur in digit ratios involving digit 5 (24). Thus, investigating sexual dimorphism in several digit ratios among adults as well as among children may be fruitful for a better understanding of these patterns and developmental changes therein.
We examined our hypotheses in two independent samples of adults (Studies I and II). This strategy of testing the same research hypotheses in more than just a single sample (i.e., not leaving replication of initial findings to other researchers and the future) constitutes an attempt of internal replication and is known as the discovery-replication sample approach. In addition, we examined the same research hypotheses in a further sample, comprised of children (Study III). All the three studies reported here received approval of the Ethics Committee of the D.A.V. College.
General Discussion
The present three studies report a pattern of sexual dimorphism from dorsally measured digit ratios similar to the pattern of sexual dimorphism reported by earlier studies for digit ratios derived from metacarpal-bone length (19). Because digit ratios derived from dorsal digit length might suitably reflect digit ratios derived from phalangeal bone length, the present evidence supports and extends the findings of Robertson et al. (12). Thus, it is conceivable that a stronger sexual dimorphism occurs in digit ratios that have digit 5 as one of the constituents in the ratios, as derived from phalangeal bone length.
Previous studies have reported patterns of noticeable ethnic and geographic differences in the 2D:4D ratio, either measured ventrally (27, 28) or from bone lengths (11). For both ventral and bone digit ratios, Europeans populations (ventral measurements: 0.985–0.994; bone measurements: 0.905–0.913) present higher 2D:4D, followed by Asian populations (ventral: 0.983–0.989; bone: 0.902–0.914), and African populations (ventral: 0.977–0.986; bone: 0.891–0.90) (11, 28). For dorsal length measurements, 2D:4D data are available for European populations [men: 0.96 (21)], which are higher than those reported for the samples from India in the present studies (0.94). Thus, the pattern of ethnic and geographic differences in 2D:4D derived from either dorsal, ventral, or bone digit length appears to be similar. The present study reports a pattern of sexual dimorphism in digit ratios similar to the one reported by McFadden and Bracht (19) for a Caucasian population. Thus, the occurrence of a relatively longer digit 5 in men than in women does not seem to depend on ethnicity or geographic region. Indeed, research suggests that, for the 2D:4D ratio, sex differences generalize across ethnic groups and world regions (11, 27, 28). It would be interesting to study the pattern of sexual dimorphism in dorsally measured digit ratios among African, European, and other populations.
The pattern of sexual dimorphism in dorsal digit ratios, as reported here, differs from the one that has been found for ventrally measured digit ratios (14–17). Because the main difference between dorsal and ventral digit lengths is the position of the (ventrally located) flexion creases, it appears conceivable that these could be the factor determining the differences between the patterns of sexual dimorphism in digit ratios derived from dorsal vs. ventral digit lengths. Furthermore, the tendency of a stronger sexual dimorphism in digit ratios of the left hand, when measured dorsally, differs from the evidence from earlier studies, measuring digit length ventrally, and reporting stronger sexual dimorphism in digit ratios of the right hand (1, 2, 15–17, 20, 29). A further line of studies has reported that the stronger sexual dimorphism in the 2D:4D ratio of the right hand only occurs for ventral measurements (1, 2, 29), whereas not for digit ratios derived from bone length (10, 12, 13).
The occurrence of similar patterns of sex differences in digit ratios, as measured dorsally, among children and adults supports the early determination of these sex differences. Studies have reported that the sexual dimorphic effect in the relative length of digits (2D:4D) might be related to prenatal sex-hormone levels (1, 2). However, the exact mechanism of this effect is not yet clear. Some theorists have further suggested that the ratio of prenatal androgens (testosterone) to estrogens (estradiol), i.e., the T:E ratio, may determine sexual dimorphism in digit ratios (30).
Based on a rodent model, Zheng and Cohn (31) suggested that sex differences in 2D:4D are determined by the opposing effects of androgen receptor and estrogen receptor-α activity in bone progenitor tissues of digit 4. Polymorphism in the estrogen receptor-α gene accounted for some variance (11%) in digit ratios of birds (32), although there is evidence that the epidermis only has androgen receptors, but not estrogen receptors-α (33). Moreover, testosterone injection experiments in primates during early gestational age have evidenced male-typical changes in dermatoglyphic variables (34) and in ventral, but not bone, 2D:4D ratios (35). It is therefore possible that prenatal testosterone has effects on both the bones and the epidermis, whereas estrogen effects are limited to the bones. Hence, genetic effects of estrogen receptors, together with prenatal testosterone levels, could determine sex differences in bone digit ratios, whereas the relative placing of flexion creases could depend on the latter factor, but not on the former one.
Studies have shown that the formation of the finger flexion creases likely is determined genetically (36). Because a geometric continuity (from anterior to posterior) is visible in the placing of flexion creases in a hand, the placing of flexion creases in a digit seems to be the part of a hand-wide developmental plan. Although a sexual dimorphism in the placing of flexion creases (ratio of ventral digit 3 length by hand length) has been reported (37), as of yet the relative placing of digit flexion creases itself has not been investigated. Similar to flexion creases, the relative length of fingers (i.e., digit ratios) also appears to follow a hand-wide developmental plan. To recap, the present studies show a pattern of sexual dimorphism in dorsally measured digit ratios (namely, a stronger effect in the relative length of digit 5), different from the pattern of sexual dimorphism in ventrally measured digit ratios (namely, a stronger effect in the relative length of digit 2). Accordingly, two different developmental pathways, acting within the same organ, might account for this differentiated pattern. In this context, future research would benefit from investigating both the relative placing of digit flexion creases and the relative length of digits, as measured dorsally, in order to delineate the factors determining these more clearly.
Across the current Studies I–III, sex differences in dorsally measured 2D:4D nominally were not statistically significant, although effect sizes were present (i.e., non-null) in some tests. Limited sample size (i.e., N < 150; similar to our samples) definitely accounts for such failures to reach conventional significance criteria in null-hypothesis significance testing, even when the size of effects is as expected. Several prior studies of comparable research also yielded nominally not significant sex effects in 2D:4D derived from bone digit length [(19, 23, 38); exception (39)], whereas larger samples usually yield findings that are statistically significant, even when effect sizes are of similar size [N = 327 (10); N = 1,060 (11); N = 3,172 (12); N = 250 (13)].
Furthermore, sex differences in the 2D:4D ratio have been reported to be larger for ventral than for bone measurements [for a meta-analysis, see Ref. (20)]. It thus appears that soft tissue, or the placing of flexion creases, or both, may be factors contributing to larger sex effects observed for ventrally measured digit ratios, as compared to bone digit ratios. It is interesting to note that some studies even reported reversed trends for bone digit ratios [i.e., higher male than female means (40)]. Seen in this perspective, the pattern of sex effects in 2D:4D in the current Studies I–III does not appear to be atypical.
In our data, the strongest sex effect among children (Study III) occurred for the 3D:5D ratio, whereas among adults (Studies I and II) for the 4D:5D ratio. This either may be due to chance (i.e., sampling error) or, alternatively, may point to additional interaction effects of age and sex on relative digit lengths. To replicate and clarify these findings remains an agenda for future research. In this context, it is informative to note that prior research has not yielded consistent patterns with regard to the magnitude of the sexual dimorphism seen in the various digit ratios (11, 14–17). Relatedly, relationships between sex-hormone levels and the 2D:4D ratio appear to be weak or absent [for a meta-analysis, see Ref. (41)]. Thus, it is likely that factors other than sex hormones impact on and confound the sexual dimorphism in digit ratios.
Dorsal digit length seems to be a suitable proxy for phalangeal bone length. A study of digit ratios derived from phalangeal bone length, including digit ratios involving digit 5, may further validate the present findings. In this context, investigations including individuals affected by endocrinologically informative conditions (such as congenital adrenal hyperplasia, complete androgen insensitivity syndrome, or Klinefelter syndrome) would be particularly interesting.
In conclusion, the evidence from the present series of studies suggests that digit ratios, when measured dorsally and involving digit 5 as a constituent of the ratio, present stronger sex differences than the widely investigated, classic, ventrally measured 2D:4D ratio.