Digit ratio predicts sporting success among female fencers independent from physical experience and

Scand J Med Sci Sports 2010: 20: 853–860 doi: 10.1111/j.1600-0838.2009.01031.x

& 2009 John Wiley & Sons A/S

Digit ratio (2D:4D) predicts sporting success among female fencers independent from physical, experience, and personality factors M. Voracek, B. Reimer, S. G. Dressler Department of Basic Psychological Research, School of Psychology, University of Vienna, Vienna, Austria

Corresponding author: Martin Voracek, Department of Basic Psychological Research, School of Psychology, University of Vienna, Liebiggasse 5, Rm 03-46, A-1010 Vienna, Austria. Tel: 143 1 4277-47846, Fax: 143 1 4277-47849. E-mail: martin.voracek@univie.ac.at Accepted for publication 13 August 2009

Research particularly focusing on male athletes and popular sports (running and soccer) suggests associations of lower (masculinized) second-to-fourth digit ratio (2D:4D), a putative marker of prenatal androgen action, with better sports performance. Studies focusing on women, non-mainstream sports, or controlling for covariates relevant for sporting success are still sparse. This study examined associations between 2D:4D and performance of both male and female athletes active in fencing (a non-mainstream sport dominated by male participants), while controlling for covariates. National fencing rankings and 2D:4D of 58 male and 41 female Austrian tournament fencers (mean age 24 years) were correlated. Among female,

but not male, fencers, lower 2D:4D was related to better national fencing rankings. 2D:4D still accounted for incremental variance (12%) in fencing success, when the effects of salient performance factors (age, body mass index, years of fencing, training intensity, and the personality variables achievement, control, harm avoidance, and social potency) were controlled for (totaling 35% attributable variance).

Athletes active in the most aggressive form (the sabre) had lower 2D:4D than those active in the other forms (e´pe´e and foil fencing). Sporting success in adult life might be partly prenatally programmed via long-lasting extragenital effects of testosterone.

The second-to-fourth digit ratio (2D:4D) is a widely studied putative marker for the permanent (organizing) masculinization effects of prenatal androgen action on the human brain, physique, and behavior (Manning et al., 1998; Voracek & Loibl, 2009). Males on average display lower 2D:4D than females, with the size of this sex difference being about 0.5 standard deviation units or smaller (Manning, 2002a).

Sex and individual differences in 2D:4D are established in utero (Malas et al., 2006; Galis et al., in press) and seem sufficiently stable during postnatal growth (McIntyre et al., 2005; Trivers et al., 2006).

Men’s 2D:4D is a correlate of genetically based differential sensitivity to testosterone (namely functional polymorphisms in the X-linked androgen receptor gene; Manning et al., 2003a, b), and a higher testosterone-to-estradiol ratio in the amniotic fluid of pregnant women is associated with a lower (masculinized) 2D:4D in their subsequently born children (Lutchmaya et al., 2004).

Testosterone has many extragenital developmental effects (Bardin & Catterall, 1981). Specifically, high fetal androgen levels promote the development of efficient cardiovascular systems, good visuospatial abilities, physical endurance and speed (Manning & Taylor, 2001), and a propensity for rough-and-tum-

ble play. For this latter effect, there is abundant evidence from human research (Collaer & Hines, 1995) as well as from non-human primate species (Wallen, 1996). Given the links of the above features with sporting aptitude, a number of recent studies (bibliography: Voracek & Loibl, 2009) have investigated possible associations between 2D:4D and performance and achievement in sport.

2D:4D is lower in high school students who opted to enroll for physical education courses than in those who did not (Tlauka et al., 2008) and in female elite athletes (international and national champions; Pokrywka et al., 2005) as well as in female amateur athletes (Latourelle et al., 2008), relative to women not engaged in sports. 2D:4D is also lower in professional soccer players, relative to local population controls, internationals vs non-internationals, top-league vs lower-league players, and in players of first-team squads, compared with reserves or youthteam members (Manning & Taylor, 2001; Manning et al., 2003a, b).

Lower (masculinized) 2D:4D is correlated with better performance in male (Manning & Taylor, 2001) and female (Paul et al., 2006) amateurs across a variety of sports, faster short-distance running times in young soccer talents (Manning, 2008) and

853

Voracek et al. sprinting speed in boys (Manning & Hill, 2009), faster middle-distance running times in men (Manning, 2002a), better performance in cross-country races in men (Manning et al., 2003a, b) and in both sexes (Manning et al., 2007), higher training frequency for cross-country races (Manning et al., 2007), faster slalom skiing times (Manning, 2002b), higher self-rated performance in amateur soccer players (Manning et al., 2003a, b), better ratings of coaches for football ability in boys (Manning, 2008), and higher numbers of international appearances among professional soccers players (Manning & Taylor, 2001). Lower 2D:4D is also associated with better physical education school grades among both boys and girls (Ho¨nekopp et al., 2006), better physical fitness among both young men and women (Ho¨nekopp et al., 2006), higher circulating testosterone levels in young male swimmers (Ju¨rima¨e et al., 2008), with upper-arm and shoulder-girdle strength (shoulder-press weight lifting) in anabolic-free young men (even with controls for body weight; Manning, 2008), and with hand-grip strength in men (Fink et al., 2006), but not in women (van Anders et al., 2006; Gallup et al., 2007). There appear to be few null findings in this literature (Barone, 2009), thus indicating that associations between 2D:4D and sports performance might be relatively robust and thus are fairly consistently found. Of particular interest is the finding of Tester and Campbell (2007), that associations of lower 2D:4D and better performance across a variety of sports among amateurs of both sexes are preserved, even when several physical, experience, and personality factors of salience for sport performance are controlled for. With its inclusion of covariates from three domains salient for sport performance (namely, physique, past experience, as well as personality), the Tester and Campbell (2007) study is the only of its kind in this area, thus significantly extending the related accounts reviewed here and serving as a guiding principle for further inquiry on the topic.

The martial arts are an obvious candidate sport form for investigations of 2D:4D as a predictor of sport performance. Up to now, two studies have investigated 2D:4D in fencers (Voracek et al., 2006; Besco´s et al., 2009). Modern sport fencing essentially is a stylized fight against a striking opponent. Therefore, among all martial arts, it is probably the one that most closely mimics actual dyadic fighting with arms (such as in a duel) or past forms of warfare (lethal battles), whereas boxing or wrestling would most closely mimic unarmed physical fighting. In particular, sport fencing strongly emphasizes cardiovascular fitness, speed, manual dexterity, and short reaction times.

Key findings of the first study on 2D:4D and fencing (Voracek et al., 2006) included that, in male

854

fencers, a lower right-minus-left difference in digit ratios (DR L), which is considered as an additional pointer for high prenatal androgen levels that shows lower values in men than in women (Manning, 2002a), was associated with better current as well as highest past national fencings rankings. This effect was independent of training intensity and years of fencing. These findings were based on 37 male fencers, whereas the female subsample (17 women) was too small to expect reliable effects. Also, 2D:4D was somewhat lower in male athletes who competed in the most aggressive form (sabre fencing) than in those competing in the two other forms (e´pe´e and foil) combined. Sabre fencing is distinguished by feint and strong offense. Major differences between e´pe´e or foil vs sabre fencing are that in the latter form, the blade is also used (i.e., the sabre is not only a thrusting weapon, but also a striking weapon), and that the sabre fencing mask has a metallic covering (because the head is a valid target area for striking actions with the sabre). Female participation rates in sabre fencing frequently are even lower than in e´pe´e or foil fencing, and for these combined reasons, the sabre is commonly considered as the most aggressive form of modern sport fencing (see Voracek et al., 2006). Key findings of the second study (Besco´s et al., 2009), which utilized a multinational, all-female sample of 87 world-class e´pe´e fencers, included that lower (masculinized) right-hand 2D:4D (but not lefthand 2D:4D or DR L) corresponded to better current and highest past world rankings. Of note, this effect was even strengthened with appropriate controls for known factors relevant for 2D:4D (ethnicity) and for fencing, world rankings (years of international experience, height, and weight).

2D:4D studies pertaining to the performance of female athletes still are in the minority. Exceptions from this are the studies of Pokrywka et al. (2005) and Besco´s et al. (2009). The first, small-sample study on 2D:4D and fencing (Voracek et al., 2006) lacked statistical power to ascertain hypothesized effects in female fencers. Further, the intriguing finding of the Tester and Campbell (2007), namely that 2D:4D explains incremental variance in amateurs’ sport performance over and above those explained by physical, experience, and personality factors, has not yet been replicated, nor has it been generalized to non-amateurs. More generally, it has been noted that many findings of 2D:4D research still await replication (Putz et al., 2004). Of note, there has been a number of replication failures of original findings in this area recently, which in turn have led to concerns of possible publication bias in this field (e.g., Vehmas et al., 2006; Voracek & Stieger, 2009).

While it is true that a slight majority of the literature on associations of 2D:4D with sport performance has been published independently from Manning (11 out

2D:4D and fencing of the 20 references reviewed above), it is also true that there indeed have been few replication attempts within this specific literature, even within the set of papers not coauthored by Manning.

The above considerations combined led to the design of the present study. Its aim was to investigate the associations of 2D:4D and sporting success in both male and female fencers anew. Specifically, to increase statistical power, larger subsamples of male and female fencers than the one of the prior study (Voracek et al., 2006) were sought for. Further, the pioneering and fruitful approach of Tester and Campbell (2007) was incorporated in the design, i.e., it was tested whether digit ratios still explain variance in fencing performance, when salient factors (physique, experience, and personality) for performance have been partialled out.

Methods Sample A new, ethnically homogeneous sample (94% native Austrians), comprised of 58 male and 41 female fencers, was collected (not overlapping with the Voracek et al., 2006, study). Participants volunteered to participate in this research (i.e., they were not paid or provided with incentives). Participants’ mean age was 24 years (SD 5 11). Sample breakdown by fencing form was 41 e´pe´e, 33 foil, and 25 sabre fencers, with broadly comparable proportions of women in each category (44%, 42%, and 36% female e´pe´e, foil, and sabre fencers). For achieving this near-balance, female sabre fencers had to be slightly oversampled. Out of 99 fencers, 11 (11.1%) were lefthanded (men: 8.6%; women: 14.6%).

Procedure Data were collected during three fencing tournaments held in Austria. Athletes were approached on site and, after providing informed consent, on site completed a survey form comprising basic personal information (their name, sex, age, height, weight, fencing form, current fencing age class, training intensity, years of fencing experience, and current and highest past national fencing rankings) and the personality measures.

Subsequently, digital scan files of athletes’ hands were taken.

As the study solely involved data collection on noninvasive measures from normal, healthy adults who volunteered to participate, in accordance with current Austrian laws, ethical review and approval was not applicable. Participants were treated in accordance with the principles laid down in the Declaration of Helsinki; specifically, study participation was on an informed consent basis and under conditions of privacy and confidentiality.

National fencing rankings, taken as objective measures for fencing performance, were verified through cross-checks with official data (Austrian Fencing Association). When rankings for more than one fencing age class were stated, these were aggregated to yield a single estimate. Usage of national rankings for specific age classes already implied some correction for differences in athletes’ age and fencing experience.

Current national rankings were available for 48 male and 33 female fencers and highest past national rankings for 54 males and 33 female fencers. Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). Significance was

set to Po0.05 (two-tailed) and Po0.10 was additionally registered as a trend result.

Materials

Personality measures Following the procedures of Tester and Campbell (2007), four deliberately selected scales (Social Potency, Achievement, Control, and Harm Avoidance) from the Multidimensional Personality Questionnaire (MPQ) (Tellegen, unpublished observation; Patrick et al., 2002) were administered in its German form (Angleitner et al., 1993). Within the MPQ personality model, the first two dimensions belong to the higher-order factor Positive Emotionality, whereas the latter two to the higher-order factor Constraint. When projected onto the factorial space of other widely accepted personality models, Social Potency and Achievement surface as positive facets of Extraversion, Achievement in addition as a negative facet of Psychoticism, and Harm Avoidance and Control as negative facets of Impulsivity (Hennig & Netter, 2005). The four MPQ scales are comprised of 12 items each, with the response format being either a forced choice (agree vs disagree) to one statement or a forced choice between two opposite statements (which one is more disliked). Item responses are recoded and summed up accordingly to yield scale scores. In the above order of scales, sample reliabilities (Cronbach a) were 0.75, 0.72, 0.71, and 0.68, tallied to reference values (Tellegen, unpublished observation; Angleitner et al., 1993; Patrick et al., 2002), and item-total correlations invariably were positive.

Finger-length measures Following standard practice of 2D:4D research (Voracek et al., 2007), palmar-view, flatbed-scanned images of participants’ right and left hands were produced. From high-resolution laser printouts of these images, finger lengths were measured three times from the midline of the flexion crease most proximal to the palm to the fingertip by experienced investigators (blind to each other’s measurements) with digital vernier calipers measuring to 0.01 mm. Interobserver measurement repeatability (assessed with single-score intraclass correlation coefficients, ICCs; two-way mixed-effects model with absolute-agreement definition; Voracek et al., 2007) was found to be excellent (all Pso0.001). The ICCs were 0.999 for length of index and ring fingers of the right and left hand (R2D, R4D, L2D, and L4D), and 0.991 and 0.994 for right-hand and left-hand digit ratio (R2D:4D and L2D:4D), respectively. The threefold measurements were averaged before calculating digit ratios. Directional asymmetry in digit ratios was calculated as DR L 5 R2D:4D – L2D:4D (ICC 5 0.981, Po0.001).

Results Normative findings and validity checks Descriptive statistics for the study variables by athletes’ sex, along with tests for sex differences, are shown in Table 1. Male and female fencers were of comparable age. Self-reported training intensity and years of fencing did not differ between the sexes, whereas normative sex effects in height, weight, and BMI emerged in the sample. Female fencers had significantly higher Social Potency and Harm Avoidance scores than male fencers. Although sex differences in digit ratios were not nominally significant

855

Voracek et al.

Table 1. Sex differences in study variables

Age (years) Height (m) Weight (kg) Body-mass index (kg/m2) Training intensity (h/wk) Fencing experience (years) R2D:4D L2D:4D DR L MPQ Social Potency MPQ Achievement MPQ Control MPQ Harm Avoidance

Males (n 5 58) Mean (SD)

Females (n 5 41) Mean (SD)

t

P

d

24.2 1.78 71.2 22.4 6.6 10.0 0.959 0.952 0.007 6.36 7.60 7.31 6.28

23.6 1.67 58.3 20.9 5.7 7.6 0.971 0.959 0.012 7.63 8.07 6.51 7.78

0.24 7.76 5.83 2.83 1.12 1.37 1.75 1.03 0.88 2.09 0.86 1.48 2.77

0.81 o 0.001 o 0.001 0.006 0.27 0.18 0.08 0.31 0.38 0.04 0.39 0.14 0.007

0.05 1.57 1.19 0.50 0.24 0.29 0.38 0.22 0.21 0.43 0.18 0.30 0.56

(11.1) (0.07) (11.5) (3.1) (4.0) (9.3) (0.034) (0.034) (0.024) (3.33) (2.85) (2.42) (2.70)

(11.1) (0.07) (9.9) (2.9) (3.2) (6.6) (0.028) (0.030) (0.024) (2.40) (2.37) (2.94) (2.60)

t, test statistic from independent-groups t tests, with associated P value; d, Cohen’s d effect-size metric (male minus female mean, divided by square root of weighted mean of group variances).

with this sample size and two-tailed tests, men, as expected, had lower digit ratios than women, and the small-to-medium size of this effect was consistent with prior research (Manning, 2002a). Individuals’ R2D:4D and L2D:4D were significantly positively related (men: Pearson’s r 5 0.74, women: r 5 0.65; P’so0.001).

Also as expected, there were substantial positive correlations (P’so0.001) between athletes’ age and years of fencing (r 5 0.57), height and weight (r 5 0.68), and current and highest past fencing rankings (Spearman’s rs 5 0.87). BMI was more strongly associated with weight (r 5 0.88, Po.001) than with height (r 5 0.25, Po0.05). For both sexes, all three digit ratio measures were unrelated to age, height, weight, BMI, the four personality measures, and the two experience variables (training intensity and years of fencing), as were the latter two variables to each other, and for both male and female fencers as well as for the sexes combined, handedness was unrelated to fencing rankings (details omitted).

Replicating prior findings (Voracek et al., 2006), sabre fencers (n 5 25) had lower digit ratios than e´pe´e and foil fencers combined (n 5 74): for R2D:4D, means (and SD) were 0.955 (0.037) vs 0.967 (0.029), t(97) 5 1.62 (P 5 0.11, d 5 0.38); and for L2D:4D, 0.943 (0.035) vs 0.959 (0.031), t(97) 5 2.23 (P 5 0.03, d 5 0.50).

Digit ratios and fencing performance: bivariate associations Correlations of digit ratios with fencing rankings are shown in Table 2. Overall, associations were small and directionally inconsistent for men, whereas somewhat larger and, of importance, directionally consistent with expectation, for women. Specifically, lower (masculinized) R2D:4D in female fencers cor-

856

responded to better current as well as highest past national fencing rankings, which effects approached significance in rank-order correlational analysis (rs 5 0.330 and 0.339, respectively; two-tailed P’s 5 0.06 and 0.05).

Similar to the study of Tester and Campbell (2007), the performance indicators (rankings) had ordinal scale level. Comparisons of parametric vs nonparametric correlational analyses (Pearson’s r vs Spearman’s rs), as reported in Table 2, showed that the results obtained were directionally wholly consistent and magnitudinally broadly comparable. If anything, Pearson’s correlations were, for the most part, somewhat smaller, not larger, than corresponding Spearman’s correlations, thus in particular ruling out the presence of outliers in the data. In other words, effects appeared not to be inflated through usage of Pearson’s r, and this parametric analysis apparently worked against, not for, the research hypothesis. This pattern of results appeared justification enough to replicate the parametric multivariate analysis (reported in the next subsection) used by Tester and Campbell (2007) under a general scheme of things that was identical to theirs.

Digit ratios and fencing performance: multivariate associations Following Tester and Campbell (2007), a hierarchical multiple regression model was specified to predict fencing rankings from physical, experience, and personality variables, and from digit ratios. Physical factors (age and BMI), experience variables (training intensity and years of fencing), and the four personality measures were consecutively entered as blocks 1 to 3 into the model. Bivariate correlational analysis (above subsection) was suggestive for a lack of associations between digit ratios and fencing rank-

2D:4D and fencing ings for men, but not for women. Therefore, following a standard practice of moderated regression (Jaccard & Turrisi, 2003), to test for this interaction effect, a sex digit ratio product term (coding: men 5 1, women 5 2) was added as the last block

to the model. The results of the hierarchical regression analysis, with current rankings as the criterion, are summarized in Table 3.

Across the sexes combined, the regression model was significant from block 2 onwards. According to the final (fourth) model, almost one-half of the variance in current fencing rankings (R2 5 0.468) was attributable to the predictors. Regression diagnostics yielded no evidence for multicollinearity problems of the final model (all variance inflation factors, VIFs,o3). Across the sexes combined, BMI and personality factors did not contribute significantly to the model. More experience (as indicated by higher training intensity and more years of fencing) as well as younger age all contributed uniquely and significantly to the model. Importantly, the sex R2D:4D interaction, added in the fourth and last block to the model, explained a significant increment of variance in fencings rankings (DR2 5 0.115) over and above those explained by the preceding predictor blocks, with the sign of the

Table 2. Associations of digit ratios with national fencing rankings

Males

Females

Current ranking Best ranking Current ranking Best ranking R2D:4D

0.088 0.067 L2D:4D 0.025 0.052 DR L 0.148 0.137

0.141 0.062 0.204 0.122 0.102 0.145

0.244 0.330a 0.145 0.176 0.108 0.156

0.216 0.339b 0.100 0.177 0.127 0.191

First-line entries are Pearson’s correlations (r), second-line entries Spearman’s rank-order correlations (rs). a P 5 0.06, b P 5 0.05 (two-tailed).

Table 3. Hierarchical regression of current fencing rankings on physical, experience, and personality factors, and on digit ratios

Predictor

Model 1 (F2,78)

Model 2 (F4,76)

Model 3 (F8,72)

sr

Model 4 (F9,71)

b

sr

Age

0.185

0.167

0.334*

0.203

0.301z

0.179

BMI

0.089

0.080

0.227*

0.198

0.236*

0.199

b

sr

b

Training intensity

0.417***

0.381

0.399***

0.350

Fencing experience

0.448**

0.300

0.459**

0.299

0.123

0.106

0.114

0.107

0.072

0.061

0.066

0.059

MPQ Social Potency MPQ Achievement MPQ Control MPQ Harm Avoidance Sex 2D:4D

R2

0.056

DR2

F

2.330

DF max. VIF

1.227

0.330

0.352

0.274

0.022

9.377***

4.891 ***

15.555***

0.602

2.705

2.837

r

b

sr

0.408** 0.399** 0.079 0.080 0.409*** 0.404*** 0.519*** 0.514*** 0.044 0.047 0.102 0.108 0.091 0.088 0.154 0.156 0.395*** 0.395***

0.238 0.234 0.062 0.063 0.359 0.354 0.336 0.333 0.038 0.040 0.096 0.101 0.077 0.075 0.135 0.137 0.340 0.340

0.223 0.169 0.440 0.075 0.199 0.122 0.163 0.120 0.300 0.307

0.468 0.468 0.115 0.116 6.928*** 6.930*** 15.395*** 15.409*** 2.936 2.920

b, standardized regression coefficient; sr, semipartial correlation; r, zero-order (Pearson) correlation; DR2 and DF, change in R2 and F statistics; max. VIF, largest variance inflation factor.

For Model 4, first-line entries are for R2D:4D, second-line entries for L2D:4D. z Po0.10, *Po0.05, **Po0.01, ***Po0.001.

857

Voracek et al. associated b coefficient indicating that lower R2D:4D corresponded to better rankings only among women, whereas not among men. This significant interaction effect was almost identical for L2D:4D (DR2 5 0.116).

Rerunning the hierarchical regression model for highest past fencing rankings (details omitted) yielded very similar results (R2 5 0.459, for both R2D:4D and L2D:4D), except that lower BMI was marginally associated with better rankings and that the contributions of digit ratios to women’s, but not men’s, rankings, over and above those attributable to all other predictors, were only marginally significant and magnitudinally smaller (DR2 5 0.022). The contributions of DR L to rankings in a further hierarchical regression model were insignificant (details omitted).

Discussion This is the first study of associations between 2D:4D and sporting ability in a sample of non-amateurs from one sport form that controlled 2D:4D effects on sporting performance for a set of salient factors for sporting performance, including physical, experience, as well as personality variables. 2D:4D still accounted for incremental variance (12%) in fencing success, when the effects of physical, experience, and personality variables on sporting performance (totaling 35% attributable variance) were controlled for.

The data indicated that athletes’ age, training intensity, and years of experience were highly predictive of achieved fencing rankings. Given these facts, the independent and incremental explanative contributions of digit ratios to fencing success are intriguing.

BMI did not contribute significantly to the final model (Table 3, Model 4). This was not surprising, as among sports people, BMI strongly depends on muscle mass. Accordingly, in the present sample the variance observed in BMI was rather low (Table 1), as could be expected for a rather homogeneous group of non-amateurs. With the benefit of hindsight, the utility of additionally accounting for BMI in this research context might therefore be questioned. Nonetheless, to be on the safe side, BMI was incorporated as a covariate in the present analysis, therein following standard practice of prior related research (e.g., Manning & Hill, 2009).

As indicated by significant sex 2D:4D product terms in hierarchical regression models, the predictive effect of digit ratios for fencing rankings applied for female, but not male, fencers. In common with the literature, lower (masculinized) digit ratios were associated with better fencing rankings (Voracek et al., 2006). Commensurate results were obtained when using either R2D:4D or L2D:4D as the pre-

858

dictor and either current or highest past national fencing rankings as the criterion.

In this effects of digit ratios on fencing success were absent among male athletes and directional asymmetry (DR L) of digit ratios did not predict fencing rankings in either sex, two findings from a prior study on 2D:4D in fencers (Voracek et al., 2006) were not replicated. On the other side, the current findings of associations between 2D:4D and fencing success in female fencers from one country and across the three categories of sport fencing replicated and extended a similar finding of Besco´s et al. (2009), which was based on a multinational sample of female world-class fencers in one fencing category (the e´pe´e).

Nonreplication of initial findings and inconsistent evidence across studies are not uncommon in 2D:4D research (Putz et al., 2004; Vehmas et al., 2006). They highlight the subtlety of 2D:4D effects on index traits, which do not emerge in every sample, and the necessity to acquire sufficiently large samples in this line of research in general. With 37 male and only 17 female fencers, the prior study’s sample was not only half the size of the present one (58 male and 41 female fencers), but also had a more skewed sex ratio than the current sample. Such circumstances may well contribute to cross-study inconsistencies of some findings. Admittedly, the sample of the present study still was not as large as one would optimally desire, but samples of generous size likely cannot be hoped for in researching non-mainstream sports with rather small numbers of participants, such as sport fencing is. At any rate, the necessarily limited sample of this study constitutes one study limitation. However, two further findings from the Voracek et al.

(2006) study fully replicated with the present data: athletes active in the most aggressive form of modern sport fencing (the sabre) presented lower 2D:4D than those active in the other forms (e´pe´e and foil fencing), and left-handedness was more frequent among tournament fencers (women: 14.6%) than in the corresponding (Austrian) general population (about 5–7%; Voracek et al., 2006). The latter result also replicated a well-known peculiarity of non-amateurs of interactive sports in general and fencers in particular (Harris, in press).

Several further results, including normative findings, also attest to the typicality of the data analyzed here: repeatability of finger-length measurements and digit ratios was high, individuals’ R2D:4D and L2D:4D were strongly positively correlated, personality measures were internally reliable, and physical (age and BMI) and experience (training intensity and years of fencing) variables were meaningfully and, for the most part, substantially related to fencing achievement. Also, men had lower digit ratios than women. Although this sex difference was not nomin-

2D:4D and fencing ally significant with the current sample, which necessarily was of limited size, the sex effect observed in 2D:4D was of small-to-medium size and thus was commensurate to the one typically found in human digit ratio studies (e.g., Manning, 2002a, 2008; Voracek et al., 2006). Furthermore, the main analyses either were within-sex analyses (Table 2) or, in another form, accounted for participant sex (Table 3), so the lack of nominally significant differences in 2D:4D between the sexes could not impact these analyses. All in all, these observations add to the plausibility of the key findings.

The data were collected on site, at national fencing tournaments, and thus only a brief survey could be implemented. As a future research agenda, it would be particularly interesting to elucidate whether and how further laterality aspects (such as congruent vs crossed hand-eye and hand-foot preference; Voracek et al., 2006) relate to digit ratios and fencing success.

One further limitation of this study (and of investigations of 2D:4D and sporting ability in general) is that, by their design and the indicators used for sporting performance, they are retrospective and merely observational. Such designs are known to be prone to various sources of bias, particularly to selection effects in the samples under study. To appropriately address the question of whether it is actually possible to identify future top athletes among beginners and young talents of one sport,

prospective designs will be essential. This study used modern sport fencing as a candidate sport form for investigating prenatal programming effects of androgens, as probed with 2D:4D, on sport performance.

Fencing is an interactive, aggressive, and clearly male-type sport form (i.e., characterized by a high male-to-female ratio among its participants). It will be informative to continue this approach to similar forms of male-type sports, especially pertaining to the analysis of associations between digit ratio and sporting success among female athletes.

Perspectives On the whole, the current data add to accumulating evidence for a role of prenatal programming of sporting aptitude via sex-hormonal effects (Manning, 2002a, 2008). Sporting success in adult life might be partly promoted by early, long-lasting extragenital effects of testosterone (Collaer & Hines, 1995; Manning & Taylor, 2001), and these effects appear to be independent from the other, more proximate, factors (i.e., physical, experience, and personality) for sporting performance.

Key words: digit ratio (2D:4D), prenatal testosterone, sex differences, fencing, Tellegen Multidimensional Personality Questionnaire (MPQ)

References Angleitner A, Langert R, Schilling J, Spinath F. German form of the Multidimensional Personality Questionnaire (2nd, rev. version).

Unpublished manuscript, University of Bielefeld, Germany, 1993.

Bardin CW, Catterall JF. Testosterone: a major determinant of extragenital sexual dimorphism. Science 1981: 211: 1285–1294.

Barone LM. Digit ratio (2D:4D) and ponderal index as predictors of halfmarathon performance [abstract]. Am J Phys Anthropol 2009: 138(S48): 83.

Besco´s R, Esteve M, Porta J, Mateu M, Irurtia A, Voracek M. Prenatal programming of sporting success: associations of digit ratio (2D:4D), a putative marker for prenatal androgen action, with world rankings in female fencers. J Sports Sci 2009: 27: 625–632.

Collaer ML, Hines M. Human behavioral sex differences: a role for gonadal hormones during early development?

Psychol Bull 1995: 118: 55–107.

Fink B, Thanzami V, Seydel H, Manning JT. Digit ratio and hand-grip strength

in German and Mizos men: crosscultural evidence for an organizing effect of prenatal testosterone on strength. Am J Hum Biol 2006: 18: 776–782.

Galis F, Ten Broek CMA, Van Dongen S, Wijnaendts LCD. Sexual dimorphism in the prenatal digit ratio (2D:4D).

Arch Sex Behav, in press. doi: 10.1007/ s10508-009-9485-7.

Gallup AC, White DD, Gallup GG. Handgrip strength predicts sexual behavior, body morphology, and aggression in male college students.

Evol Hum Behav 2007: 28: 423–429.

Harris LJ. In fencing, what gives left-handers the edge? Some views from the present and the distant past. Laterality, in press. doi: 10.1080/ 13576500701650430.

Hennig J, Netter P., eds. Biopsychologische Grundlagen der Perso¨nlichkeit [Biopsychological bases of personality].

M u¨ nchen: Elsevier, 2005.

Ho¨nekopp J, Manning JT, Mu¨ller C. Digit ratio (2D:4D) and physical fitness in males and females: evidence for

effects of prenatal androgens on sexually selected traits. Horm Behav 2006: 49: 545–549.

Jaccard J, Turrisi R. Interaction effects in multiple regression, 2nd edn. Thousand Oaks, CA: Sage, 2003.

Ju¨rima¨e T, Voracek M, Ju¨rima¨e J, La¨tt E, Haljaste K, Saar M, Purge P. Relationships between finger-length ratios, ghrelin, leptin, IGF axis, and sex steroids in young male and female swimmers. Eur J Appl Physiol 2008: 104: 523–539.

Latourelle SM, Elwess NL, Elwess JM. Finger forecasting: a pointer to athletic prowess in women – a preliminary investigation by an undergraduate biology class. Am Biol Teach 2008: 70: 411–414.

Lutchmaya S, Baron-Cohen S, Raggatt P, Knickmeyer R, Manning JT. 2nd to 4th digit ratios, fetal testosterone and estradiol. Early Hum Dev 2004: 77: 23– 28.

Malas MA, Dogan S, Evcil EH, Desdicioglu K. Fetal development of the hand, digits and digit ratio

859

Voracek et al.

(2D:4D). Early Hum Dev 2006: 82: 469–475.

Manning JT. Digit ratio: a pointer to fertility, behavior, and health. New Brunswick, NJ: Rutgers University Press, 2002a.

Manning JT. The ratio of 2nd to 4th digit length and performance in skiing. J Sports Med Phys Fitness 2002b: 42: 446–450.

Manning JT. The finger book: sex, behaviour and disease revealed in the fingers. London: Faber & Faber, 2008.

Manning JT, Bundred PE, Newton DJ, Flanagan BF. The second to fourth digit ratio and variation in the androgen receptor gene. Evol Hum Behav 2003a: 24: 399–405.

Manning JT, Bundred PE, Taylor R. The ratio of 2nd and 4th digit length: a prenatal correlate of ability in sport. In: Reilly T, Marfell-Jones M, eds.

Kinanthropometry VIII: proceedings of the 12th Commonwealth International Sport Conference.

London: Routledge, 2003b: 165–174.

Manning JT, Hill MR. Digit ratio (2D:4D) and sprinting speed in boys. Am J Hum Biol 2009: 21: 210–213.

Manning JT, Morris L, Caswell N. Endurance running and digit ratio (2D:4D): implications for fetal testosterone effects on running speed and vascular health. Am J Hum Biol 2007: 19: 416–421.

Manning JT, Scutt D, Wilson J, LewisJones DI. The ratio of 2nd to 4th digit length: a predictor of sperm numbers and concentrations of testosterone, luteinizing hormone and oestrogen.

Hum Reprod 1998: 13: 3000–3004.

860

Manning JT, Taylor RP. 2nd to 4th digit ratio and male ability in sport: implications for sexual selection in humans. Evol Hum Behav 2001: 22: 61–69.

McIntyre MH, Ellison PT, Lieberman DE, Demerath E, Towne B. The development of sex differences in digital formula from infancy in the Fels Longitudinal Study. Proc Biol Sci 2005: 272: 1473–1479.

Patrick CJ, Curtin JJ, Tellegen A. Development and validation of a brief form of the multidimensional personality questionnaire. Psychol Assess 2002: 14: 150–163.

Paul SN, Kato BS, Hunkin JL, Vivekanandan S, Spector TD. The big finger: the second to fourth digit ratio is a predictor of sporting ability in women. Br J Sports Med 2006: 40: 981–983.

Pokrywka L, Rachon˜ D, SucheckaRachon˜ K, Bitel L. The second to fourth digit ratio in elite and non-elite female athletes. Am J Hum Biol 2005: 17: 796–800.

Putz DA, Gaulin SJC, Sporter RJ, McBurney DH. Sex hormones and finger length: what does 2D:4D indicate? Evol Hum Behav 2004: 25: 182–199.

Tester N, Campbell A. Sporting achievement: what is the contribution of digit ratio? J Pers 2007: 75: 663–678.

Tlauka M, Williams J, Williamson P. Spatial ability in secondary school students: intra-sex differences based on self-selection for physical education. Br J Psychol 2008: 99: 150–156.

Trivers R, Manning JT, Jacobson A. A longitudinal study of digit ratio

(2D:4D) and other finger ratios in Jamaican children. Horm Behav 2006: 49: 150–156. van Anders SM, Vernon PA, Wilbur CJ. Finger-length ratios show evidence of prenatal hormone-transfer between opposite-sex twins. Horm Behav 2006: 49: 315–319.

Vehmas T, Solovieva S, Leino-Arjas P. Radiographic 2D:4D index in females: no relation to anthropometric, behavioural, nutritional, healthrelated, occupational or fertility variables. J Negat Results Biomed 2006: 5: 12.

Voracek M, Loibl LM. Scientometric analysis and bibliography of digit ratio (2D:4D) research, 1998– 2008. Psychol Rep 2009: 104: 922–956.

Voracek M, Manning JT, Dressler SG. Repeatability and interobserver error of digit ratio (2D:4D) measurements made by experts. Am J Hum Biol 2007: 19: 142–146.

Voracek M, Reimer B, Ertl C, Dressler SG. Digit ratio (2D:4D), lateral preferences, and performance in fencing. Percept Mot Skills 2006: 103: 427–446.

Voracek M, Stieger S. Replicated nil associations of digit ratio (2D:4D) and absolute finger lengths with implicit and explicit measures of aggression. Psicothema 2009: 21: 382–389.

Wallen K. Nature needs nurture: the interaction of hormonal and social influences on the development of behavioral sex differences in rhesus monkeys. Horm Behav 1996: 30: 364–378.

Copyright of Scandinavian Journal of Medicine & Science in Sports is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.