CC BY-NC-ND
0Egyptian Journal of Forensic Sciences (2014) 4, 7-16
Contents lists available at ScienceDirect
Egyptian Journal of Forensic Sciences
journal homepage: www.ejfs.org
Egyptian Journal of Forensic sciences
ORIGINAL ARTICLE
Stature estimation from footprint measurements in Indian Tamils by regression analysis
T. Nataraja Moorthy a,% Ahmad Mustaqqim Bin Mostapa a, R. Boominathan N. Raman b
Forensic Science Program, University Sains Malaysia, Kubang Kerian, Kelantan, Malaysia VHNSN College, Madurai Kamaraj University, Virudhunagar, India
Received 5 June 2013; revised 22 September 2013; accepted 29 October 2013 Available online 5 December 2013
KEYWORDS
Forensic science; Forensic anthropology; Stature estimation; Footprint; Indian Tamils
Abstract Stature estimation is of particular interest to forensic scientists for its importance in human identification. Footprint is one piece of valuable physical evidence encountered at crime scenes and its identification can facilitate narrowing down the suspects and establishing the identity of the criminals. Analysis of footprints helps in estimation of an individual's stature because of the existence of the strong correlation between footprint and height. Foot impressions are still found at crime scenes, since offenders often tend to remove their footwear either to avoid noise or to gain a better grip in climbing walls, etc., while entering or exiting. In Asian countries like India, there are people who still have the habit of walking barefoot. The present study aims to estimate the stature in a sample of 2,040 bilateral footprints collected from 1,020 healthy adult male Indian Tamils, an ethnic group in Tamilnadu State, India, who consented to participate in the study and who range in age from 19 to 42 years old; this study will help to generate population-specific equations using a simple linear regression statistical method. All footprint lengths exhibit a statistically positive significant correlation with stature (p-value < 0.01) and the correlation coefficient (r) ranges from 0.546 to 0.578. The accuracy of the regression equations was verified by comparing the estimated stature with the actual stature. Regression equations derived in this research can be used to estimate stature from the complete or even partial footprints among Indian Tamils.
© 2013 Production and hosting by Elsevier B.V. on behalf of Forensic Medicine Authority.
* Corresponding author. Tel.: +60 97677589/129224610 (HP); fax: + 60 97657515.
E-mail address: nataraja@kb.usm.my (T. Nataraja Moorthy). Peer review under responsibility of Forensic Medicine Authority.
1. Introduction
Every part of the body is different in its own way not only within a particular body, but also from other bodies. There is also a relationship between each part of the body and the whole body. Nothing exemplifies this truth more than the relationship that various parts of the body have to the stature of an individual.1 In this manner, an individual's footprint may represent his or her identity. Person identification using foot-
2090-536X © 2013 Production and hosting by Elsevier B.V. on behalf of Forensic Medicine Authority. http://dx.doi.Org/10.1016/j.ejfs.2013.10.002
prints is also an emerging biometric technique.2 Footprints as valuable physical evidence available at crime scenes are used to link the crime to the perpetrator. Footprints can be collected from almost all types of crime scenes, and the possibility of their recovery at the scenes of sexual offenses and homicides is relatively more.3 Characteristic features may provide useful clues to establishing personal identity whenever complete or partial footprints are recovered at the crime scene and that can help in including or excluding the possible presence of an individual at the scene of the crime.4
Examination of barefoot impressions is important, especially in Asian countries like India where the majority of the rural population walk barefoot. This is largely due to socioeconomic and climatic conditions. The partial or complete footprints can be found on rain covered surfaces; newly waxed floors; freshly cemented surfaces; moistened surfaces; in dust, mud, sand, oil, paint; and can be left in blood at the murder scenes3'5'6 An aspect of human identification that has received scant attention from forensic anthropologists is the study of human feet7 and the footprints8 made by the feet. Researchers have been conducting stature estimation research using foot9-26 and footprints27-33 because of the existence of a strong correlation between one's stature and foot size34 From the analysis of footprints, definitive information on many physical characteristics of the individuals who made them can be retrieved. The information on footprint (and foot) morphology is especially significant because it elucidates the individuality of each person's footprints.8'11
Many of these studies were conducted for stature estimation from foot and footprints on mixed populations. The researchers have cautioned that the people from different regions of a country bear different morphological features depending upon their geographical distribution and primary racial characteristics; therefore, a single formula cannot represent all parts of that country or world.3'30'35-38
For the stature estimation from foot/footprint parameters' the researchers concluded that toes to heel length measurements in a foot/footprint are more reliable and accurate than from any other measurements like breadth at ball/heel' big toe breadth/ length' etc.3'12'14'18-20'22 This present study aims to estimate the stature from various footprint length measurements in Indian Tamils' an ethnic group in Tamilnadu State' South India. The linear regression analysis method was used for stature estimation since the reliability and prediction of stature estimation are more accurate and reliable with the regression analysis method.38 The investigation revealed that the correlations of stature with various footprint length measurements from different toes to heel in both left and right are high. The accuracy of regression equations was verified by comparing the estimated stature with actual stature. The estimated values are found close to the actual stature values. Scatter graphs were drawn by plotting various footprint length measurements against statures and scatter graphs analysis showed perfective positive correlation forming an elliptical pattern of the distribution of values' i.e. best fit lines were obtained in all graphs.
2. Materials and methods
The present study was conducted on consented adult male Indian Tamils' an ethnic group residing in Tamilnadu state' South India. Fig. 1 below depicts the site location' i.e. sampling
point of this research. This study aimed to estimate stature in a sample of 2'040 bilateral footprints collected from 1'020 adult male Tamil volunteers of ages ranging between 19 and 42 years old. The research procedure followed was in accordance with the approved ethical standards of Universiti Sains Malaysia Research Ethics Committee (Human). Before the sample collection' information such as subjects' name' age' and place of origin was obtained and recorded.
Those with any apparent disease' orthopedic deformity' injury or disorder were excluded from the study. The subjects were confirmed to be descendants from three generations of Tamils to ensure no genetic variation within races that can disrupt the results as stature can be affected by not only environment' but also genetic makeup. Just prior to research participation' the subjects were advised to wash their feet with soap and water. A cleaned plain glass plate of 8 mm thickness was uniformly smeared with ''Kores quick drying black duplicating ink 4746'' with the help of a footprint roller. The subject was asked to step with the left foot on an inked glass plate with minimal pressure. Then the inked foot was placed on an A4 plain white paper kept aside on a uniform surface and thus the left footprint was transferred. Before lifting the sole from the paper' anatomical landmarks of the feet were marked on the papers close to the footprints which are mid-rear heel point and most anterior point of all toes. Following Robbins27 and Krishan3' the designated longitudinal axis (DLA) and baseline (BL) were drawn on the footprints. The DLA is from the pter-nion (P) landmark at the mid-rear heel margin to the lateral side of the toe 1 pad margin' the axial line touches the rim of the pad margin as it passes forward beyond the length of foot. Base line (BL) is drawn at the rear edge of the foot and perpendicular to the DLA. The base line extends from the landmark P at the rear of the heel in both medial lateral directions while maintaining its perpendicular alignment with the DLA. Its axis can be determined as marked on the footprint using the pro-tractor. With the 90L mark on the footprint placed on the DLA' and the midpoint of the protractor base at pternion' one automatically has the perpendicular BL by drawing a line through the pternion along the base of the protractor. Then five diagonal footprint length measurements were taken from the mid-rear heel point (P) to the most anterior point of each left toe (LT1 LT2' LT3' LT4' and LT5). The left footprint length measurements were designated as PLT1' PLT2' PLT3' PLT4' and PLT5. The procedure was repeated for the right footprint and the right footprint length measurements were designated as PRT1 PRT2' PRT3' PRT4' and PRT5. The landmarks and diagonal length measurements on the left footprint are shown in Fig. 2. The following are the diagonal length measurements taken on the left and right footprint of each male participant. All footprints and information relating to participants were coded with sample ID for anonymity.
2.1. Left and right footprint length measurements
i. PLT1 - length' measurement taken from the mid-rear heel point' pternion (P) to the most anterior point (LT1) of toe 1 on the left footprint.
ii. PLT2 - length' measurement taken from the mid-rear heel point' pternion (P) to the most anterior point (LT2) of toe 2 on the left footprint.
iii. PLT3 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (LT3) of toe 3 on the left footprint.
iv. PLT4 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (LT4) of toe 4 on the left footprint.
v. PLT5 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (LT5) of toe 5 on the left footprint.
vi. PRT1 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (RT1) of toe 1 on the right footprint.
vii. PRT2 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (RT2) of toe 2 on the right footprint.
viii. PRT3 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (RT3) of toe 3 on the right footprint.
Figure 1 Map of India showing the sampling point, Tamilnadu State, South India. (Source: http://www.google.com.my/ search?q = india + map + states).
ix. PRT4 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (RT4) of toe 4 on the right footprint.
x. PRT5 - length, measurement taken from the mid-rear heel point, pternion (P) to the most anterior point (RT5) of toe 5 on the right footprint.
2.2. Stature measurement
Stature of each subject was measured in cm nearest to mm by Seca 213 portable stadiometer following the standard proce-dure3 Stature is the vertical distance between the vertex and the floor when the individual was standing barefooted with
Figure 2 Landmarks and diagonal length measurements on left footprint PLT1-PLT5, length, measurements taken from the mid-rear heel point, pternion (P) to the most anterior point of toes LT1-LT5 on left footprint. DLA, designated longitudinal axis.
Figure 3 Scatter graphs showing strong positive correlation between various footprint length measurements (PLT1-PLT5, PRT1-PRT5) and stature PLT1-PLT5, length, measurements taken from the mid-rear heel point, pternion (P) to the most anterior point of toes LT1-LT5 on left footprint. PRT1-PRT5, length, measurements taken from the mid-rear heel point, pternion (P) to the most anterior point of toes RT1-RT5 on right footprint.
Table 1 Descriptive statistics of footprint length, stature and frequencies of fifth toes contact (PLT5, PRT5) in adult male Indian Tamils.
Variables N Range Max-Min (cm) Min (cm) Max (cm) Mean (cm) SD
Stature 1020 38.10 152.00 190.10 173.69 4.57
PLT1 1020 6.80 21.30 28.10 24.72 1.02
PLT2 1020 6.60 21.20 27.80 24.63 1.06
PLT3 1020 6.70 20.30 27.00 23.69 1.04
PLT4 1020 6.50 19.30 25.80 22.41 0.97
PLT5 990* 6.40 17.60 24.00 20.67 0.92
PRT1 1020 6.10 21.40 27.50 24.62 1.01
PRT2 1020 6.40 21.20 27.60 24.52 1.06
PRT3 1020 6.40 20.30 26.70 23.60 1.04
PRT4 1020 6.30 19.40 25.70 22.32 0.98
PRT5 994* 6.00 17.50 23.50 20.59 0.90
Min, minimum. Max, maximum. PLT1-PLT5, left footprint lengths from anterior point of toes LT1-LT5 to mid-rear heel point P. PRT1-PRT5, right lengths from anterior point of toes RT1-RT5 to mid-rear heel point P. SD, standard deviation. Fifth toes contact (PLT5, contact: 990, non-contact: 30) (PRT5, contact: 994, non-contact: 26).
Table 2 Means, standard deviations and values of't' of bilateral difference (left-right) in footprint measurements in adult male Indian
Tamils.
Measurements (cm) N Mean bilateral difference (left-right) S.D t-value*
T-1 (PLT1-PRT1) 1020 0.10 0.43 7.264
T-2 (PLT2-PRT2) 1020 0.11 0.42 8.027
T-3 (PLT3-PRT3) 1020 0.09 0.40 7.682
T-4 (PLT4-PRT4) 1020 0.09 0.40 6.781
T-5 (PLT5-PRT5) 994 0.08 2.97 3.713
PLT1-PLT5, left footprint lengths from anterior point of toes LT1-LT5 to mid-rear heel point P. PRT1-PRT5, right footprint lengths from
anterior point of toes RT1-RT5 to mid-rear heel point P. N, number of participants. SD, standard deviation.
¿-Values are significant at P < 0.01 level.
the head held in the Frankfurt horizontal plane with eyes looking forward. The feet were positioned side by side in a standing position, arms by the sides. Heels, buttocks and upper back were in contact with the wall when the measurement was made. Considering the diurnal variation in stature, the heights of the subjects were measured approximately at the same time in the evening. The diurnal change in height of a person was indicated as early as 1726 and the shortening in stature during daytime was reported and confirmed by the researchers.39,40 All the measurements were taken by the author (TN) to avoid inter-observer error. Each participant's stature was measured four times with the support of Forensic Assistant Directors. The first time the participant was invited to get into the stadiometer, the stature was noted by the main author. The participant was then requested to get down from the stadiometer. The subject was asked again to get into the stadiometer for recording the stature. Thus the procedure was repeated four times and finally recorded the concordant stature value in centimeters.
The data were analyzed using PASW Statistics version 18 (Predictive Analytic Software). Bilateral asymmetry was calculated for each of the footprint measurements and tested for significance using one sample t-test. Pearson's correlation coefficients between various footprint length measurements and stature were obtained. The linear regression analysis method was employed to derive regression equations for stature estimation from various footprint length measurements since stature estimation from footprint length is more accurate and reliable with regression analysis.38 Regression equations are an algebraic expression of regression lines (the linear relationship between two variables) represented by a straight line. The data were analyzed by using scatter graphs drawn by plotting footprint lengths against stature demonstrating the relationship between the footprint length and stature as shown in Fig. 3.
3. Results
Table 1 presents the descriptive statistic of stature, left and right footprint length measurements and frequencies of fifth toe contact during the footprint developing process in all subjects. The stature ranges from 152.00 to 190.10 cm (mean 173.69 cm). First toe-heel footprint length measurement (PLT1, PRT1) is found to be the longest in both right and left sides.
All left footprint length measurements (PLT1-PLT5) were found larger than the right footprint lengths (PRT1-PRT5). Thus, it can be interpreted that since the size of the left foot is larger than the size of the right foot in the sample, this indicates statistically significant bilateral asymmetry. First toe length measurements (PLT1 and PRT1) are found to be the longest on the right and left footprints. An important finding in the investigation was that both the left footprint length and the right footprint length measurements gradually declined from first toe-heel to fifth toe-heel in left (24.7220.67 cm) footprint and right (24.62-20.59 cm) footprint. The degree of decrease between the T4 and T5 of both feet showed a higher difference compared with the difference from T1 to T3. This occurred because LT5 of 30 subjects and RT5 of 26 subjects did not make contact with the ground during the footprint development process. This similar phenomenon was observed in other footprint studies.31,33,37,41
Table 3 Linear regression equations for stature estimation through various footprint length measurements in adult male Indian Tamils (N = 1020).
Variable Regression equations SEE (cm) R R2
PLT1 113.117 + 2.450 PLT1 3.812 0.546 0.298
PLT2 112.502 + 2.485 PLT2 3.730 0.578 0.334
PLT3 113.899 + 2.524 PLT3 3.748 0.573 0.328
PLT4 114.410 + 2.645 PLT4 3.773 0.564 0.318
PLT5 116.056 + 2.789 PLT5 4.522 0.561 0.315
PRT1 112.148 + 2.499 PRT1 3.802 0.555 0.308
PRT2 112.902 + 2.479 PRT2 3.741 0.575 0.331
PRT3 115.138 + 2.481 PRT3 3.768 0.566 0.320
PRT4 115.579 + 2.603 PRT4 3.791 0.559 0.312
PRT5 114.528 + 2.875 PRT5 3.756 0.568 0.323
PLT1-PLT5, left footprint lengths from anterior point of toes LT1-LT5 to mid-rear heel point P. PRT1-PRT5, right footprint lengths from anterior point of toes RT1-RT5 to mid-real heel point P. SEE, standard error of estimate. R, Pearson's correlation coefficient. R2, the coefficient of determination. P < 0.001.
Table 2 presents the means, bilateral difference (left-right) in footprint lengths, standard deviation, p-values and t-values. All length measurements show statistically significant asymmetry and T1 and T2 lengths are found to be significantly more asymmetric in the sample. The highest t-value was found for T2 (8.027) and lowest for T5 (3.713).
Table 3 presents the linear regression equations for estimation of stature from footprint length measurements in left and right footprints. The table also shows the Karl Pearson's correlation coefficient (R) between stature and bilateral footprint length measurements for all the subjects. All the correlation coefficients show positive relationship and statistical significance (p < 0.001). The highest correlation coefficient was noted for PLT5 (0.578) and the lowest correlation coefficient noted was for PLT1 (0.546). Hence, statistically significant correlation coefficients exist between stature and all footprint length measurements. With regard to the coefficient of determination (R2), the predictive accuracy is found to be statistically significant for stature estimation.
The regression line represents the predicted score. It passes through the exact center of the data in the scatter diagram. It must be remembered that with the help of the regression equations alone, perfect prediction is practically impossible. So, an indication is needed to show how inaccurate the prediction might be, i.e., the standard error of estimate (SEE). The SEE predicts the deviations of the estimated stature from the actual stature. If the SEE is zero, then there is no variation of the regression line and the correlation is perfect. Thus, with the help of the SEE it is possible to ascertain how good and how representative the regression line as a description of the average relationship between the two series. The SEE shows a lower value while estimating the stature from all bilateral footprint length measurements. Standard errors ranged between 3.730 and 3.812 cm. The left toe 2 (PLT2) exhibits the lowest value of SEE and the left toe 1 (PLT1) exhibits the highest value of SEE in the subjects.
All scatter graphs show an elliptical pattern of distribution of values. The analysis of the scatter diagram indicated a strong positive correlation between footprint length measure-
ments (the independent variable, X) and stature (the dependent variable, Y).
4. Discussion
India is a land of enormous genetic, cultural and linguistic diversity42 Tamils or Tamil people are an ethnic group native to Tamilnadu, India. Tamilnadu, literally ''Land of Tamils,'' lies in the southernmost part of the Indian Peninsula. It is bound by the Eastern Ghats in the north, the Anamalai Hills and Palakkad on the west, by the Bay of Bengal in the east, the Gulf of Mannar in the Pak Strait in the south-east and by the Indian Ocean in the south. Tamilnadu State has been the home of the Tamil people since at least 500 BCE and the language is Tamil and has a long tradition of culture. The bordered States have different populations viz. Malayalees in Kerala State, Telugus in Andhra Pradesh State and Kannadas in Karnataka State with varying culture, climate, food habits and languages. Limited studies were conducted on footprint for stature estimation in India. The result of this study provided regression equations for the stature estimation from various bilateral footprint length measurements of Tamils in South India when complete and even partial footprints are found at the scenes of crimes for human identification.
The age of the subjects ranged between 19 and 42 years. Commonly, stature at 18 years is accepted as adult, although there are small increments in stature after this.43'44 Some researchers indicated that the foot in a male grows to its adult size by 16 years of age.45,46 Hence, the minimum age was fixed at 19 years old to conduct this study. The investigation reveals that the left footprint length measurements are found to be larger than the right footprint length measurements.
Another salient feature observed was that the mean values of the footprint length gradually declined from first toe to fifth toe in both left footprints (24.72-20.67 cm) and right footprints (24.62-20.59 cm). The degree of decline showed a big drop from T4 to T5 when compared with T1-T3 in both feet. This occurred because of the fifth toe which was found missing in some of the subjects' footprint. The size of the left foot is larger than the size of the right foot in the sample and indicated the existence of statistically significant bilateral asymmetry. This bilateral asymmetry in the lower limbs of Tamils is consistent with Fawzy's study on the Egyptian population,30
while Krishan found the asymmetry in T-2, T-4 and T-5 on the Gujjars of North India3 and Kanchan found asymmetry in T1-T3 in the Indian population.33 The mean footprint length measurements of Tamils showed appreciable size variation compared with the mean footprint lengths of other populations like North Indian Jat Sikh, North Indian Gujjar, Indian, Malaysian Malays and Egyptian. Table 4 shows the comparison of means of various footprint length measurements of adult males between the present study and other studies. The first toe-heel measurements in the present study are found to be the longest in both right and left sides (PLT1 and PRT1) which is consistent with Egypt's. The Gujjar's left second toe (PLT2) and right first toe (PRT1) are found to be the longest. The present investigation revealed that the mean footprint length measurements of Tamils are longer than Gujjars of the North Indian population3, and Malays of Malaysian population33, but shorter than the mean footprint length of Jat Sikh of the North Indian population36 and the Egyptian pop-ulation.30 Thus, the comparison of footprint length measurements clearly indicated that the morphological size of footprints formed from the feet of Indian Tamils varies considerably from the footprint size of other populations in India and outside India.
It is common knowledge that right-handed persons prefer to kick with the right foot, and this has been correlated with the dominance of the contra-lateral cerebral hemisphere.45 However, observations recorded by Singh43 and Chhibber47 suggested that in the majority of both right-handed and left-handed persons, the left lower limb is more used than the right. These observations include greater wear on the left shoe, a marked tendency to put the left foot forward first on starting to walk, and the ability to apply greater pressure with the left foot. This in turn enlarges the bones of the dominant foot and therefore, produces a footprint of larger dimensions.3 Philip did not find significant bilateral asymmetry while working on the footprints of the south Indian population.1 Robbins also did not find significant bilateral asymmetry in various measurements of the feet of the U.S. population.48 Both researchers stated that the measurements of most variables in person's left and right bare footprint are similar enough to permit either the right or left foot from being used for height and weight examinations.
Another important observation made during the development process of the footprint is that the fifth toe of some
Table 4 Comparison of mean male footprint length measurements (cm) between the present study and other studies.
Variables Present study Tamils Gujjars3 Jat Sikh36 Egyptian30 Malays37
(cm) (South India) (North India) (North India) (Egypt) (Malaysia)
PLT1 24.72 24.05 27.13 25.31 24.26
(Longest toe-heel length PLT1 or PLT2) (Longest toe-heel length PLT1 or PLT2)
PLT2 24.63 24.15 25.16
PLT3 23.69 23.45 24.27
PLT4 22.41 21.88 23.14
PLT5 20.67 20.78 21.49
PRT1 24.62 24.13 26.26 24.82 24.27
(Longest toe-heel length RT1 or RT2) (Longest toe-heel length PRT1 or PRT2)
PTR2 24.52 23.93 24.69
PRT3 23.60 23.51 23.80
PRT4 22.32 21.34 22.67
PRT5 20.59 20.09 20.94
Table 5 Comparison of frequencies in non-contact of fifth toes on ground during the development process of footprints between male Indian Tamils and other population (male) studies.
Present study Kanchan33 Nataraja Moorthy37 Sarah Reel31
Frequency % Frequency % Frequency % Frequency %
56 in 1020 5.5 4 in 50 8 10 in 113 8.8 5 in 31 16.1
Table 6 Comparison of estimated stature with actual stature from all footprint length measurements of randomly selected subjects.
No Actual Stature (cm) Estimated stature from various footprint length measurements (cm)
PLT1 PLT2 PLT3 PLT4 PLT5 PRT1 PRT2 PRT3 PRT4 PRT5 Mean
1 172.00 172.13 171.53 171.33 170.99 170.46 172.34 171.74 171.55 171.19 170.61 171.40
2 173.00 173.12 172.52 172.01 172.06 172.04 173.51 172.92 172.94 173.68 173.92 172.87
3 174.00 174.11 173.74 173.80 174.41 174.78 175.65 175.10 174.33 174.10 174.86 174.50
4 175.00 173.91 173.94 174.41 174.20 174.52 174.09 176.05 175.52 175.14 175.81 174.76
5 176.50 174.90 176.15 176.68 176.13 175.65 175.26 176.05 176.12 175.55 174.86 175.74
6 177.00 176.69 176.55 175.86 175.70 176.55 176.23 176.44 175.73 176.39 176.04 176.20
7 178.00 178.08 177.36 178.53 179.14 177.68 177.20 176.64 178.11 178.26 177.67 177.87
8 179.00 177.48 177.96 177.50 177.00 177.46 176.81 177.82 177.12 176.59 176.75 177.25
9 180.00 177.90 177.76 177.30 177.42 177.70 178.80 178.80 177.71 177.22 177.22 177.71
10 181.00 180.00 179.57 179.56 179.14 178.13 178.18 179.20 179.30 178.46 177.70 179.00
subjects was found to be missing, i.e., did not make contact with the ground and was reflected as missing toes in the footprints. The non-contact of the fifth toe is an important and valuable clue in crime scene investigation and perpetrator identification in a scientific way. Similar findings were observed and recorded by Reel et al.31, Kanchan33 and Nataraja Moorthy.37
Table 5 presents the comparison of frequencies in non-contact of fifth toe with the ground during the footprint development process between the present study and other studies. The percentage in non-contact of the fifth toe on the ground during the development process of the footprint is lower in Tamils than other studies. The intrinsic muscle function would help to stabilize the arch along with the plantar aponeurosis and help in maintaining the toes flat on the ground until lift-off has occurred.49
The footprints can be classified into four types on the basis of the relative morphological lengths of the first, second and third toes.4 These four types have been denoted as T-Type (the tibialis-type), F-type (the fibularis-type), M-type (the mid-ularis-type) and O-type (the intermediate-type). The present investigation revealed that the footprints of Indian Tamils have fallen either in T-type or F-type unlike Gujjars, the footprints of which found distributed in all types with varying frequencies.
The present study successfully derived linear regression equations for stature estimation from 10 diagonal axis footprint length measurements and the regression equations present lower SEE. The SEE values are found to be in between 3.730 and 3.812 cm. The foot length is the best parameter for estimating stature since stature can be estimated from an unknown person with great accuracy and a small SEE, i.e. about 2-6 cm.22 The second left toe (LT-2) shows a lower SEE (3.730) while the first left toe (LT1) (3.812) shows a higher SEE. Thus the stature can very well be estimated from footprint lengths with low SEE. The correlation coefficient be-
tween the various footprint length measurements with stature shows a high positive linear relationship between footprint length and stature, and the correlation coefficient ranged from 0.555 to 0.578.
The accuracy of the regression equations was verified by comparing the estimated stature with actual stature. All the 10 footprint length measurements of the 10 randomly selected subjects were substituted in the derived regression equations to estimate the stature so that the equations may be used even for partial footprints. The estimated stature values are found to be closer to actual stature values and are presented in Table 6.
The present investigation revealed that both regression equations and scatter graphs indicated the existence of statistically significant positive correlation between footprint lengths and stature of Indian Tamils. It is a usual and regular trend in a real crime scenario wherein most of the crime scenes are disturbed by the public and inmates by leaving their footprints at the crime scene area before arrival of police investigating officers. Hence it is the duty of the investigating officer to recognize and locate the appropriate footprints suitable for stature estimation so as to link effectively the crime scene and the perpetrator forensically.
5. Conclusion
The present study concludes that footprint length measurements have a strong relationship with the stature of adult male Indian Tamils. This study showed that the footprints of Indian Tamils are different from other Indian populations. It is well known that people from different regions of a country and world have different morphological features depending on their geographical distribution and racial characteristics. Hence it is important to keep it in mind that the regression equations derived in this investigation are unsuitable for any other population either in India or anywhere in the world.
Therefore, it is suggested that similar stature estimation from footprint studies should be continued for other populations living in India and in the world for the meaningful forensic investigation.
Funding
The author did not received any specific funding for this research.
Conflict of interest
The authors have no conflict of interest to declare. Ethical approval
Ethical approval was taken from Universiti Sains Malaysia.
Acknowledgments
We are thankful to all participants who took part in this strenuous study. Authors are grateful to Vice Chancellor, Dy, Vice Chancellor (research), Campus Director, Dean of Health Sciences, Forensic Science Program Chairman, Universiti Sains Malaysia for their permission and encouragement for the sample collection at Tamilnadu State, South India. Thanks are due to Mr. K. Ramanujam, I.P.S., Director General of Police, Tamilnadu State, South India for his encouragement and support to conduct this research. We wish to thank Prof. Dr. P. Sundarapandian, Prof. Dr. A. Shanmugasundaram and the Research Department of Chemistry staff and students, VHNSNC, Virudhunagar, affiliated to Madurai Kamaraj University, South India for their support in sample collection.
We are very thankful to Mr. Mani, and Mr. Donald Ravichan-dran, Assistant Directors of Forensic Sciences Department, Government of Tamilnadu, South India for their assistance throughout the sample collection.
Our thanks are due to Dr. Aniza Abd. Aziz, Biostatistics and Research Methodology Unit, and Dr. P.T. Jayaprakash, Associate Professor in Anthropology, Universiti Sains Malaysia for their help in the statistical analysis and proofreading the article in this study.
References
1. Philip TA. Formulae for establishing stature from foot size by regression method. J Ind Acad Forensic Med 1990;12:57-62.
2. Ambeth Kumar VD, Ramakrishnan M. Legacy of footprints recognition - a review. Int J Comput Appl 2011;35:9-16.
3. Krishan K. Estimation of stature from footprint and foot outline dimensions in Gujjars of north India. Forensic Sci Int 2008;175:93-101.
4. Krishan K. Individualizing characteristics of footprints in Gujjars of north India—forensic aspects. Forensic Sci Int 2007;169:137-44.
5. Qamra SR, Sharma BR, Kaila P. Naked foot marks—a preliminary study of identification factor. Forensic Sci Int 1980;16:145-52.
6. Sharma BR. Forensic Science in criminal investigation. 3rd ed. India: Universal Law Publishing Co., Pvt. Ltd; 1990.
7. Kanchan T. Somatometry of the foot in identification of dismembered remains. J South India Medicol Assoc 2010;2:56-9.
8. Robbins LM. The individuality of human footprints. J Forensic sci 1978;23:778-85.
9. Krishan K, Kanchan T, Passi N. Estimation of stature from the foot and its segments in a sub adult female population of North India. J Foot Ankle Res 2011;4(1):24.
10. Krishan K, Kanchan T, Passi N, DiMaggio JA. Stature estimation from the lengths of the growing foot-a study on North Indian adolescents. The Foot 2012;22(4):287-93.
11. Giles E, Vallandigham PH. Height estimation from foot and shoeprint length. J Forensic Sci 1991;36:1134-51.
12. Singh TS, Phookan MN. Stature and foot size in four Thai communities of Assam, India. Anthropol Anz 1993;51:349-55.
13. Anil A, Peker T, Turgut HB, Ulukent SC. An examination of the relationship between foot length, foot breadth, ball girth, height and weight of Turkish university students aged between 17 and 25. Anthropol Anz 1997;55:79-87.
14. Hilmi O, Yasemin B, Cannan D, Akin T, Mehmet E. Stature and sex estimate using foot and shoe dimensions. Forensic Sci Int 2005;147:181-4.
15. Gulsah Z, Ipek E, Zehra D. Stature and gender estimation using foot measurements. Forensic Sci Int 2008;181:54e1-5.
16. Kanchan T, Menezes RG, Moudgil R, Kaur R, Kotian MS, Garg RK. Stature estimation from foot length using universal regression formula in a north Indian population. J Forensic Sci 2010;55:163-6.
17. Jahar JK, Vijay P, Paliwal PK. Estimation of height from measurements of foot length in Haryana region. J Indian Acad Forensic Med 2010;32:231-3.
18. Rani M, Tyagi AK, Ranga VK, Rani Y, Murai A. Stature estimates from foot dimensions. J Punjab Acad Forensic Med Toxicol 2011;11:26-30.
19. Qamra SR, Jit I, Deodhar SD. A model for reconstruction of height from foot measurements in a adult population of north west India. Indian J Med Res 1980;71:77-83.
20. Sen J, Ghosh S. Estimation of stature from foot length and foot breadth among Rajbanshi: a indigenous population of north Bengal. Forensic Sci Int 2008;181:55e1-6.
21. Sonali K, Ashish R. Estimation of stature from measurement of foot length, hand length and head length in Maharashtra. Indian J Basic Appl Med Res 2012;1:77-85.
22. Krishan K, Abihilasha S. Estimation of stature from dimension of hand, feet in north Indian population. J Forensic Leg Med 2007;14:327-32.
23. Jasuja OP, Singh J, Manjari J. Estimation of stature from foot and shoe measurements by multiplication factors: a revised attempt. Forensic Sci Int 1991;50:203-15.
24. Patel SM, Shah GV, Patel SV. Estimation of height from measurements of foot length in Gujarat region. J Anat Soc India 2007;56:1-3.
25. Pawar S, Zambara BR, Sawant VG, Shinde, Bodigarry RB. Determination of personal height from foot length in Maharashtra region. Indian J Forensic Med Pathol 2011;4:59-62.
26. Mohanty BB, Agrawal D, Mishra K, Samantsinghar P, Chinara PK. Estimation of height of an individual from foot length: a study on the population of Odisha. Int J Rev Sci 2012;2:69-74.
27. Robbins LM. Estimating height and weight from size of footprints. J Forensic Sci 1986;31:143-52.
28. Hu XY, Yao HF, Lin JH. Comprehensive analysis of the correlation between the height of a person and the length of his/ her footprint. Fa Yi Xue Za Zhi 2005;21:15-8 Article in Chinese.
29. Oberoi DV, Kuruvilla A, Saralaya KM, Rajeev A, Ashok B, Nagesh KR, et al. Estimation of stature and sex from footprint length using regression formulae and standard footprint length formulae respectively. J Punjab Acad Forensic Med Toxiol 2006;6:5-8.
30. Fawzy IA, Kamal NN. Stature and body weight estimation from various footprint measurement among Egyptian population. J Forensic Sci 2010;55:884-8.
31. Reel S, Rouse S, Vernon W, Doherty P. Estimation of stature from static and dynamic footprints. Forensic Sci Int 2012;219:1-3.
32. Vidya CS, Shamsundar NM, Saraswathi G, Nanjaiah. Estimation of stature using footprint measurements. Anat Karnataka 2011;5:37-9.
33. Kanchan T, Krishan K, Shyamsundar S, Aparna KR, Jaiswal S. Analysis of footprint and its parts for stature estimation in Indian population. The Foot 2012;22:175-80.
34. Krishan K, Kanchan T. Foot length is a functional parameter for assessment of height. The Foot 2013;23:54-5.
35. Laskowski GE, Kyle VL. Barefoot impressions - a preliminary study of identification characteristics and population frequency of their morphological features. J Forensic Sci 1988;33:378-88.
36. Jasuja OP, Singh J, Jain M. Estimation of stature from foot and shoe measurements by multiplication factors: A revised attempt. Forensic Sci Int 1991;50:203-15.
37. Nataraja Moorthy T, Mazidah K, Hadzri M, Jayaprakash PT. Estimation of stature based on foot length of Malays in Malaysia. Aust J Forensic Sci 2011;43:13-26.
38. Krishan K, Kanchan T, Sharma A. Multiplication factor versus regression analysis in stature estimation from hand and foot dimensions. J Forensic Leg Med 2012;19:211-4.
39. Whitehouse RH, Tanner JM, Healy MJ. Diurnal variation in stature and sitting height in 12-14-year-old boys. Ann Hum Biol 1974;1:103-6.
40. Krishan K, Vij K. Diurnal variation of stature in three adults and one child. Anthropologist 2007;9:113-7.
41. Nataraja Moorthy T, Wan NZ, Saat M. A study on footprints of Malaysian athletes and non-athletes for application during forensic comparison. Malays J Forensic Sci 2011;2:29-35.
42. Majumder Partha P. Ethnic populations of India as seen from an evolutionary perspective. J Biosci 2001;26:533-45.
43. Singh I. Functional asymmetries in lower limbs. Acta Anat 1970;77:131-8.
44. Roche AF, Davila GH. Late adolescent growth in stature. Pediatrics 1972;50:874-80.
45. Mysorekar VR, Nadekar AN, Sarma TSR. Estimation of stature from parts of humerus and radius. Med Sci Law 1982;22:178-80.
46. Rao NK, Kotian MS. Footprint ratio (FPR) - a clue for establishing sex identity. J Ind Acad Forensic Med 1990;12:51-6.
47. Chhibber SR, Singh I. Asymmetry in muscle weight and one-sided dominance in the human lower limbs. J Anat 1970;106:553-6.
48. Robbins LM. Footprints-collection, analysis and interpretation, Charles C, Thomas, Spring Field, IL, USA, 1985.
49. Teerawat K, Sith T, Natee D. A study of footprints in athletes and non-athletic people. J Med Assoc Thai 2004;87:788-93.
