RELATIONS BETWEEN ANTHROPOMETRIC AND MOTOR VARIABLES OF SEVEN-YEAR-OLD GIRLS WITH THE DIFFERENTIATED LEVEL OF NUTRITIONAL STATUS

The objective of this paper is to examine the frequency of nutritional status and the relations between anthropometric characteristics and motor skills of the first-grade elementary schools girls from Valjevo. A battery of 16 anthropometric measurements and 9 motor tests was conducted on the pertinent sample (N = 332) of seven-year-old girls. Data were analyzed using common measures of central tendency, measures of variability and canonical correlation analysis. Based on the measured body mass index, 75% of participants have normal body mass, 21% has increased body mass, and 4% is overweight. The results showed that in the group of participants with normal body mass there is no significant connection between two systems of variables (morphological characteristics and motor skills). Data from the group of overweight participants show that the two statistically relevant correlations were extracted in the morpho-motoric space and were interpreted as canonical functions: (1) circular dimensionality of the skeleton and body mass, and explosive strength and (2) subcutaneous fatty tissue and segmented run and running speed. Canonical factors indicate that the greater the size of chest, thigh, shin and body mass, the lower are the results of the composite motor tests (plyometric jump, forward lean-backward lean-throw, long jump and standing jump), that is the girls who had higher values of back skinfold, stomach skinfold, thigh skinfold and shin skinfold, had worse result realization of motor tasks (hand-tapping, foot-tapping and running 5 x 10 m). Empirical data of the research on the isolated canonical structure are relevant from the theoretical point of view, but they also have relevant implications for identifying increased level of nutritional status and the advancement of physical education in elementary school education.


INTRODUCTION
Nutritional status of schoolgirls and its specifics represent the basic indicator of physical growth and development, which is the result of the genetic potential and interaction with the outside environment.Defining the nutritional status enables tracking growth and physical development of girls with the aim of identifying disorders or disorder risks based on the deviation level from the values recommended for the given age.Based on the current standards of the World Health Organization -WHO, the most prevailing indicator for nutritional status assessment is body mass index -BMI which represents different ratio between body height and body mass (WHO, 2012).
Increased body mass implies great amount of fatty tissue in the body, whereas obesity is determined as abnormal amount of fatty tissue which arose from misbalance of intake and outtake of energy where intake dominates which leads to organ dysfunction and deterioration of health.World Health Organization emphasizes that in the period of infancy and adolescents, the frequency of obesity tripled from the state 25 years ago, where even 10.0% of the child world population is obese or are at risk of becoming obese.This chronic endocrine metabolic non communicable disease is becoming the dominant prob-22 lem of public health in XXI century, especially in the USA but also in Europe (NCS, 2016).The main reason for the great interest of the researchers for examining this phenomenon represents the trend of the increase of obesity with children and adolescents worldwide, which the research of D'Hondt and associates shows (2014).
In his research, Bralić (2014) believes that overweight and obesity are caused by multiple interactions of hereditary and acquired factors, mainly by decrease of physical movement (exercise) and bad eating habits.In the last two decades, the number of obese children increased multiple times, so today in the world there are over 150 million obese children as the study by Cunningham, Kramerand and Venkat Naraya (2014) shows.The same authors consider that the chances for a person to be obese in adulthood are 6.2 times greater with obese children than with the ones with normal body mass.The relevant data from the developed countries at the beginning of XXI century indicate that the progressive increase in body mass occurred, and the number of obese people surpassed the number of the malnourished people in industrialized countries, where more than third of the adult population was obese, while more than 96% of the population had increased body mass, as it is stated in the study by Ogden, Carroll, Kit and Flegal (2014).
Obesity in infancy affects morbidity and mortality in adulthood, unrelated to their body mass.The research findings of Ogden, Carroll, Curtin, Lamb and Flegal (2010) stress that the number of obese children aged 6 to 11 in the period from 1980 to 2008 increased almost three times, it rose from 7% to 20%.In the countries of the EU there is a significant number of overweight and obese children, going from 13 % in the Czech Republic, 33% in Greece to 36% in Italy, and in Latin America and the countries of the west coast of the Pacific Ocean there is 12% of overweight and obese children aged 6 to 11, while 13% of the children is obese, which is shown in the research by Mendoзa, McLeod, Chen, Nicklas and Baranowski (2014).
Research findings in our country indicate that in 2000 from the entire number of adults, 54% of them were overweight -36.7% of who fall under the category of pre-obese, and 17.3% under obese (Ministry of health, Serbia, 2007).In Serb Republic there are 10.0% of underweight, 16.7% overweight, while 13.3% of preschool children are obese, as Đer-manović, Miletić and Pavlović (2016) state in their research.It has been determined on the sample of Slovenian children aged 7 to 18 that the percentage of overweight children in 1991 was 12.0%, in 2011 it was 17.2%, while the percentage of obese children in 1991 was 2.1% and in 2011. it was 5.5%, as it is stated in the study conducted by Kovač, Jurak and Leskošek (2012).
Longitudinal findings from the research by Keane, Kearney, Perry, Kelleherand and Harrington (2014) point to the fact that 30% of the children aged 7 to 14 were overweight or obese, and that the increased body mass and obesity represent the main health problem for 20-34% of the children in the Republic of Ireland.Research conducted by Mithuri and associates indicates to the trend of increase of overweight and obese children in the less developed countries of the sub-Saharan Africa, especially with girls, children in urban areas and children with higher socioeconomic status.
In their paper, Buttitta, Iliescu, Rousseau and Guerrien (2014) have determined that motor skills represent dominant factor for children's healthy physical and mental development.According to the longitudinal findings of Greier, Riechelmann and Burtscher (2014) obese children achieve lower scores in all motor tasks than the normally nourished children, and the excess of fatty tissue makes mobility and lifting one's own weight more difficult, slows down motor development and formation of motor habits.With the increase of physical activity of children within the experimental group, there is three times lesser possibility for them to later be obese than the children in control group, as it is stated in the research by Sigmund, El Ansariand and Sigmundova (2012).
The results of the study by Pereira and associates (2011) show that the Portuguese children with increased body mass, when compared to normally nourished children, have lower scores of aerobic capacity and shoulders and arms muscle mass, and the obese children aged 6 to 11have lower scores in the fitness parameters (the strength of shoulders and arms, speed and agility, strength of torso and legs, frequency of the hand movement speed, static hand strength and explosive leg strength).In their paper, Esmaeilzadeh and Ebadollahzadeh (2012) determined that obese children aged 7 to 11, when compared to normally nourished children, achieve lower scores in all motor tasks, balance, hand move-ment speed, flexibility and hand grip, while the overweight children, when compared to normally nourished children, achieve lower scores in tests: mile run, pull-up and push-up hold.
Statistically significant differences in motor tasks depend on the nutritional status of the sevenyear-old participants: girls with normal body mass achieved the best results on motor tests of explosive strength, and obese girls were more successful in the explosive hand strength; obese girls achieved the lowest and normally nourished girls the highest scores on the coordinated movement tests, while the girls with normal body mass had insignificantly better scores of the speed of movement frequency than the obese girls, which is shown in the research by Pantelić, Kostić, Đurašković, Uzunović and Ranđelović (2012).According to the research by Podstawski and Boryslawski (2012), male participants aged 7 to 9 had higher body height, body mass and BMI than the female participant, boys were more successful on motor tests of strength and girls in the tasks of flexibility.
Therefore, the matter of this research is morphological characteristics (different nutritional status) and motor skills of seven-year-old girls.Based on the main research objective, the genetic hypothesis (X 0 -linear correlation of anthropomorphic parameters and motor skills with overweight participants is statistically significant) was settled, as well as two separate hypotheses (X 1 -linear correlation of anthropomorphic parameters and motor skills with normally nourished participants is statistically significant and (X 2 -linear correlation of anthropomorphic parameters and motor skills with obese participants is statistically significant).
The participants were separated in three groups -normally nourished, overweight and obese.During anthropomorphic measuring and motor test conducted in September 2016, all the students were clinically healthy and were part of the regular phys-ical education classes.Participation in this research was voluntary.In accordance with the ethical principles of the research, school principals gave formal consent for the students' participation in the testing, which was conducted according to preaganged schedule, during regular physical education classes.

Sample of antropomorphic variables and measuring procedure
Assessment of morphological characteristics of participants was determined in accordance with the protocol of Mišigoj-Duraković (2008), battery of 16 standard anthropometric manifest variables which assess the four-dimensional space of the following fundamental latent dimensions: I) longitudinal dimensionality of the skeleton [body height (mm), leg length -spina-iliakaanteriorsuperior (cm), arm length (cm)], II) transversal dimensionality of the skeleton [shoulder width -diacromion span (cm), pelvic width -bicrystal span (cm) and hip width (cm)], III) circular dimensionality and body mass [mean chest circumference , outstretched upper arm circumference(cm), thigh circumference (cm) and shin circumference (cm) and body mass (kg) and IV)subcutaneous fatty tissue [upper arm skinfoldin the line of m. tricepsbarchi (mm), back skinfold -subscapular skinfold -(mm), stomach skinfoldsuprailiocristal(mm), thigh skinfold (mm) and shin skinfold (mm)].
Anthropometre was used to measure body height, medical weight scales -body mass, caliper using John Bull's method -skinfold, and centimeter measuring tape for body circumference.
Suitable conditions for the efficiency of statistical anthropometric measuring were enabled while complying with the predetermined norms.Anthropological instruments were examined before starting measuring, which was conducted by the same person during morning classes at the same time (± 2 h), using same instruments and technique.Before measuring, marker was used to precisely mark relevant anthropometric points and levels.During anthropomorphic measuring, barefoot participants were wearing standard gym clothes.Measuring of the even segments was conducted on the left side of the body.Themeasuring results were being read while the instrument was on participant.The measuring was conducted 3 times in order to remove any possible mistake, and then the arithmetic mean of the results was calculated.
To determine children's level of nourishment derived variable of body mass was used, that is to say limit value recommended by the World Health Organization (2016), that classified our sample using formula of the Belgian mathematician Adolphe Quetelet [BMI = body mass (kg)/(body height 2 (m) 2 ] into three nutritive categories: normally or averagely nourished (BMI = 18.5-24.9kg/m), overweight (BMI = 25-29.9kg/m 2 ) and obese (BMI ≥ 30-35 kg/m 2 ).

Sample of motor variables
Motor skills were assessed using battery with 9 measuring instruments (Metikoš, Hofman., Prot, Pintar, and Oreb, 1989; Bala and Stojanović, 2007): I) explosive strength (plyometric jump, forward lean-backward lean-throw, long jump and standing jump); II) coordination (skipping horizontal jumping rope, 20 lunges with stick thread and run and roll); III) segmented speed and running speed (hand tapping, foot tapping and running 5 x 10 m).
According to protocol, before testing, the examiner explained in detail and made task demonstration once or twice, and then did standard warm-up with children doing the five-minute run with motor tasks (changing direction, squatting and jumping) and dynamic stretching and muscle stretch.The testing was conducted by the station rules, with circular working regime, where there were at the same time 10 participants maximum wearing gym clothes, one by one doing the task three times, where the best score was used.The order of motor tests depended on the relationship between energy and informatics components which contain task suitable for neutralizing negative influence of fatigue happening after difficult movement.

Data analysis
Common measures of central tendency were used for data processing -arithmetic mean (M), variability measures -standard deviation (SD), asymmetric distribution coefficient -skewness (Sk) and shape coefficient (outside and length) data -kurtosis (Ku).Statistical relevance of the hypothesis on relation between groups of variables was tested using linear model of canonical-correlation analysis.Limit level of significance was expressed on the probability level (p ≤ .05).All results are shown in the form of tables and graphs.Data analysis was conducted using statistic package SPSS 19 for Windows.

RESULTS
Nutritional status of the participants was as follows: 75% were normal weight, 21% of over-weight patients, a 4% obese.Therefore, most girls belong to the group of normally nourished, ¼ is overweight, and the smallest number of them is obese Table 1 shows common descriptive statistical indicators of morphological and motor variables with normally nourished, overweight and obese first grade elementary school girls.Annotation:M= arithmetic mean; SD = standard deviation;Sk = standardized skewness(asymmetry coeficient -distribution distortion or skewness); Ku = standardized kurtosis (coefficient of outside and length distribution).
Analyzing the indicators of statistical distortion within the group of participants with normal body mass, it can be seen that the standard deviation of the great number of examined morphological variables is less than 1/3 of arithmetic mean of the results, which indicates to minimal differences, that is to say variability of the arithmetic mean original results, as well as to good statistical characteristic -delicacy (discrimination), successful discrimination of the large number of participants concerning measuring and adjusting, which shows the study by Bulajic, Jeremić and Radojičić (2012).However, relevant deviations of empirical values from arithmetic mean are evident with the following anthropometric characteristics: upper arm skinfold (SD = 4.89), back skinfold (SD = 5.01), stomach skinfold (SD = 6.03), thigh skinfold (SD = 6.57), shin skinfold (SD = 6.08), as well as composite motor tests: plyometric jump (SD = 6.06), skipping horizontal jumping rope (SD = 2.48) and forward lean-backward lean-throw(SD = 15.75).
Checking deviation of the attained arithmetic mean distribution in the second group of overweight participants the following standard deviations of three variables appeared: shin skinfold (SD = 6.88) and plyometric jump (SD = 4.96), skipping horizontal jumping rope (SD = 1.37), which statistically deviate a lot from arithmetic mean, while in the group of obese participants data of five variables show statistically significant deviations from arithmetic mean: back skinfold (SD = 7.40), stomach skinfold (SD = 8.42), shin skinfold (SD = 3.88), plyometric jump (SD = 3.99) and skipping horizontal jumping rope (SD = 1.66).
Attained division values of the results for assessing distribution skewness in the group of participants with normal body mass suggest that the wrong frequency with most applied variables statistically relevantly does not deviate from theoretical model of normal division of the distribution density function because asymmetry measurements lean toward zero.Most of the skewness values are formed around mean value which is in the range of good and acceptable limit values.That means that the discriminatory measurements are good, the tests are not too difficult or too easy but they suit the population of seven-year=old girls, which the research by Kolmaev and Kalinin (2016) shows.
However, testing the normality of score distribution in the group of overweight participants, positive asymmetry was determined in anthropometric and motor variables: shoulder width (Sк = 1.08), chest circumference (Sк = 1.14) and body mass (Sк = 1.28), forward lean-backward lean-throw(Sк = 1.79), skipping horizontal jumping rope (Sк = 1.18) and run and roll (Sк = 1.16).Also, in the category of obese participants it can be seen that the skewness values deviate from normal distribution of anthropometric variables: back skinfold (Sк = 1.56), stomach skinfold (Sк = 1.63),20 lunges with stick thread (Sк = 1.24), running 5 x 10 m (Sк = 1.15).According to study by Bubić (2015), attained empirical values of the right side result distribution indicate that the result distribution curve leans towards lower values, that is to say it has lower scores than the one of normal distribution.In addition, variables do not possess good delicacy (discrimination), because there was probably some bias during measuring, for example influence of systemic factors on the attained results, or the motor tasks for assessing movement coordination and running speed were too difficult due to low muscle strength of the participants and were not, due to difficulty, suitable to conduct.
Evaluation of the distribution homogeneity was done using common kurtosis coefficient.Distribution of the attained kurtosis values suggests that most of the empirical distributions do not statistically significantly deviate from Gaussian (mesocratic) bell curve, which points to homogeneity of morpho-mo-tor variable distribution on the participant sample.
Calculated positive predicators of kurtosis coefficient with overweight and obese participants point to violation of normal distribution in this measured variables, that is to say they point to blurry or leptokurtic distribution of empirical values.Characteristics of this distribution are: the arrangement is more elongated than the normal one, the tails are "deeper" than the ones from Gaussian distribution, homogeneity of the results distribution is increased due to pronounced concentration of results around arithmetic mean.

Correlation between morphological characteristics and motor skills
In order to examine maximal linear connection between manifested morpho-motor variables from two group from he normally nourished sample in Table 3., canonical correlation analysis was used, which revealed that extracted canonical functions re not statistically significant and will not therefore be used in interpretation since there is, due to extremely low mutual variance, no connection between analyzed system of variables.Also, the size of the subsample of obese participants did not, due to homeostasis phenomenon, meet the criterion for using multivariate parametric linear model, which the study by Popov and Sotnikov (2016) shows.Annotation: Rc = canonical correlation; R 2 = square canonic correlation (coefficient of determination); χ 2 = Barthel's Chi-square test ; df = degree of freedom; p = level of statistical significance To determine the nature of relation in morpho-motoric space of manifested morpho-motoric variables with overweight participants, the method of canonical correlation analysis in Table 3 extracted two pairs of statistically significant canonical factors (paired variants).While interpreting the results of this two-dimensional structure only two canonical correlations will be used, because the results of-Barthel's Chi-square test of the remaining correlated orthogonal canonical functions did not manifest as statistically significant because of the mutual sharing of the entire variability, as Mihitarяn (2016) states.
First statistically significant canonical correlation (Rc = .68)encompasses 29% of the explained mutual variance proportion, and the second (Rc= .61)12% of the mutual variability of total empirical results variance of the two systems of variables, which with statistical significance of 5% points to good overlap of the measuring items of the applied measuring instruments.On the other hand, the remaining 71%, which is 88% of the residual, remaining or unexplained score of the squared deviation from the regression line is the consequence of systematic influence of uncontrolled and unexamined variables which are not included in this research, which should not, according to the research by Sazin and Ivanova (2014), be interpreted because of the accidental oscillations and possible influence of other latent dimensions which are not included in this research.Annotation: Rc = canonical correlation; R 2 = square canonic correlation (coefficient of determination); χ 2 = Barthel's Chi-square test ; df = degree of freedom; p = level of statistical significance Table 4 shows canonical coefficients (ponders) or loads on the sample of overweight participants which are used to consider partial contributions of manifested variables on the formed canonical functions.Annotation: Constitution of the first significant and interpretable pairs of linear composite canonical functions in the morphological space is maximally determined by the positive and relatively low values of the canonical structural variable coefficient of the chest circumference (r = .46),thigh circumference (r = .42),shin circumference (r = .39)and body mass (r = 45).The attained canonical mean saturation in the matrix points to the fact that these synthetic variables are valid representatives of extracted latent dimension.Taking into account the fact that the partial intensity of every manifested variable in the morphological space higher than 0.30 and its load on the canonical function (root), structure of linear combination in the first pair of canonical factors is mostly hypothetically defined by variables which evaluate circular dimensionality of the skeleton and body mass.On the other hand, the values of standardized canonical coefficients of the remaining variables of this unipolar isolated canonical dimension are statistically insignificant.
The matrix of the linear combination structure (function) of the second pair of canonical factors projects dominantly four variable of weak intensity positive linear correlations: back skinfold (r = .39),stomach skinfold (r = .36)and thigh skinfold (r = .33)and shin skinfold (r = .31).Based on the relative involvement of each manifested variable in the pair of linear composites which represent latent construct (factor), the structure of the second canonical factor can theoretically be interpreted as subcutaneous fatty tissue.
Table 5 shows the canonical load in motor space on the subsample of overweight participants.Considering the results of the mutually orthogonal two-dimensional latent structure of motor space, it can be seen that the frame of the first and most relevant pair of canonical factors determines the negative and relatively weak linear correlations of three variables: plyometric jump (r = -.67),forward lean-backward lean-throw(r = -.53) and standing long jump (r = -.49), while the values of standardized canonical coefficients of the remaining variables will not be explored since there is no statistically significant linear interaction within the participants' motor behavior.According to maximal saturation projections, this unipolar canonical factor can hypothetically be identified as explosive strength since it maximally activates muscle cells within time unit.
Vectors of variables which have dominant projection of negative sign in motor space on linear combinations of the second pair of extracted canonical factors are: hand tapping (r =-.51), foot tapping (r = -.57) and running 5 x 10 m (r = -.55), while the projections of the remaining variables are insignificant and will not be explained.Attained moderate canonical loads signal that these synthetic variables are good representatives of isolated latent dimension of participants' motor behavior.Taking into account attained maximal values of structural correlation coefficient, this pair of linear composites, which represent latent construct (factor), can be theoretically determined as canonic factor speed.

DISCUSSION
The level of nourishment with seven-year-old girls represents relevant indicator of their health status, psychophysical skills and the assumed possibility of healthy growth and physical development.Any significant variations from standard body mass point to current or potential symptoms of health disorder or disease.Current data from Europe and around the world concerning modern man disease coincide with our research findings because it is found that 75% of participants has normal body mass, around one fifth is overweight, whereas 4% of the girls is obese.
Based on the standardized descriptive parameters analysis, it can be concluded that the groups of normally nourished, overweight and obese first grade elementary school students differ in almost all morphological criteria.The participants who have normal weight have minimal parameters of every variable of anthropometric space.Maximal values of latent variable of longitudinal dimensionality of the skeleton are present with overweight girls, whereas obese girls have maximal values of latent variables of transversal dimensionality of the skeleton, volume,body mass and subcutaneous fatty tissue.
The measured values of anthropometric measures of central tendency and variability of our research for the most part coincide with the results of the empirical works by Đorđević (2015), Đorđević  Attained descriptive parameters in motor space point to differences between groups of participants with different nutritional status.The best results in latent dimension explosive strength while measuring the strength of leg extensors, that is to say plyometric -eccentric contraction, were achieved by the normally nourished participants, then overweight, and the worst results were achieved by obese girls.The worst results and negative influence of leg triceps on motor tests plyometric jump and standing long jump for assessing velocity of leg triceps strength is the consequence of increased body mass with overweight and obese girls, it is stated in the research by Arianna Ceschia and associates (2016).On the other hand, in the latent dimension of arm and torso explosive strength, the best results of short-term maximal muscle force mobilization were achieved by obese girls, then overweight, while normally nourished girls achieved the lowest results.From the biomechanics point of view, the increased exit velocity of 3kg medical ball on motor tests forward lean-backward leanthrow can probably be explained by the longer height and arm length of obese and overweight girls than the height and arm length of their normally nourished peers.
The differences between descriptive parameters of participants with different nutritional status are also present in latent dimension (in construct) of frequency of arm and leg movement speed.Normally nourished girls, compared with overweight and obese girls, touched in turns more times with fingers the board, or the drawn square on the wall.The attained empirical findings on the reduced speed level of separate movements, or the change of body position from one to the other spot in the same line, are fully consistent with the findings from the empirical works of Chagas, Daniel, Carvalho, JoyceBatista and Luiз (2016).
Normally nourished girls from our sample achieved, compared to overweight and obese girls, better results on motor tests of latent dimension explosive leg strength, because the groups of participants with their increased body mass and subcutaneous fatty tissue lifted and carried, with considerably more difficulty, body mass in space.This finding is partly consistent with research results of Ostojić, Stojanović, Stojanović, Marić and NJardi (2011).
The linear model of canonical correlation analysis used on this sample of normally nourished girls did not isolate statistically significant canonical factors and therefore was not interpreted.Besides, relatively minimal number of participants does not enable the application of multivariate method of data processing due to parsimony or wrong conclusion.
In the isolated canonical dimension movement coordination, the best results were achieved by normal-ly nourished girls, than overweight, while the worst results in precise and fast arm and leg movement and spatial orientation were achieved in the group of obese participants, which coincides with the findings of Lopes, Maia, Rodrigues and Malina (2012); Lopes, Stodden, Bianchi, Maia and Rodrigues (2012).
While assessing the latent dimension speed, relatively uniform results in subgroups of normally nourished and overweight girls were found, where the overweight girls achieved the best result in the frequency of arm and leg movement speed and change-of-direction running, and the worst results were achieved by the obese participants.On the other hand, with these latent dimensions the physiological basis which consists of mechanism of impulse transfer speed control and regulation, the most successful were normally nourished girls, then overweight, while the worst results were achieved by obese girls.
Based on the calculated statistically relevant canonical structural correlation coefficient and the explained percentage of mutual variance of the total variability in morpho-motor space of overweight participants, it can be concluded that, with the conclusion error of 1%, the derivation of canonical function (root) is justified.Its latent structure or the construct in the canonical correlation matrix consists of mutually linearly connected two pairs of statistically significant factors (pair variants), which contain maximal proportion of the mutual variability percentage.The content of redundant bipolar structure of the vector variables of the first pair of canonical factors linear combinations in anthropomorphic space is interpreted as circular dimensionality of the skeleton and body mass, and the content of the other structure of vector variables of the second pair of canonical factor in motor space as explosive strength.
The determined relations (maximal canonical correlation and proportion of the percentage of explained mutual variance shared by canonical variables and canonical load) in morphological and motor space with bipolar first pair of canonical factors point to the very good intensity of canonical correlation and mutual variance.This points to, with significance (p ≤ .05)that the noticed linear interaction between the attained canonical functions formed by accident, or that the participants who have bigger canonical loads with variable of circular dimensionality of the skeleton and body mass -bigger chest circumference, thigh circumference, shin circumference and body mass, probably achieve worse results in the latent dimension of explosive strength (composite motor tests for assessing leg extensor strength and explosive strength of the torso, shoulders and arms), that is to say in plyometric jump, forward lean-backward leanthrow, and long standing jump for evaluating velocity strength of leg extensors.
Latent function of canonical saturations of the statistically significant, but weaker second bipolar pair of canonical factor's linear combinations, which is independent from the first pair of canonical variants, is defined as subcutaneous fatty tissue and segmented speed and running speed.Relations of this canonical function (root) that reflect part of the variance which manifested variable shares with new canonical function point to the fact that participants who have bigger values of back skinfold, stomach skinfold, thigh skinfold and shin skinfold, probably achieve worse lower results of composite movement structures: hand tapping, foot tapping and running 5 x 10 m, due to negative signs of this time variables which indicate reverse scaling.Conducting general analysis of the statistically significant canonical structure, which is independent from the first extracted pair of canonical variants, it is concluded, with huge probability, that the predictor (subcutaneous fatty tissue), with maximal correlation between orthogonal linear combinations (canonical) of the original variables and proportion of the determined mutual variance percentage, represents interfering latent dimension or construct in motor task performance which demands frequency of hand and foot movement and change-of-footdirection running speed with first grade elementary school students.The nature of determined canonical relations in the population of seven-year-old girls points to high intensity connection of the examined morpho-motor variables, that is partly identical to the model described in the studies by Đorđević, Pantelić, Kostić and Uzunović (2014).
Analyzing the attained data from the conducted transversal research, couple of significant Methodological limitations (small pertinent sample of participants, gathering data in a very short period of time, demographic markings of the analyzed sample, limited age, assessing only female participants and correlative nature of the study) were noticed.Therefore, the possibility of interpreting causal relationship and generalizing attained empirical results in morpho-motor domain is partly limited, which diminishes the validity of the conclusions about the population of sevenyear-old girls.
With the aim of overcoming the limitations of this research and giving more valid conclusions about the somatic status, latent morpho-motor structure and relations between two groups of variables, future researches should: a) use experimental research design which will enable drawing conclusion about causal relations between the assessed constructs, as well as longitudinal models which will enable determination of the development process and the changes over time, b) enable various data generators about dependent and independent variables and include time difference between measuring predictive and criteria variables, c) include more representative accidental samples of different gender and age categories from various cultures in Serbia, d) also use other variables for early detection of overweight and obese participants, for example waist circumference -WC for determining abdominal obesity, fatty tissue distribution index (Waist circumference -WHR), and e) balance methodology so that the attained results can be compared to other empirical studies.

CONCLUSION
Basic question asked in this quantitative research was: can the skills of the first grade elementary school female students on different nutritional level be predicted based on morphological characteristics.
Comparing empirical findings of our research and discussion, it is possible to formulate following conclusion with the probability level of 95%.
1. Data on different nutritional status based on BMI show that the three quarters of participants are normally nourished, around one fifth is overweight and 4% is obese.
2. The general hypothesis tested, that the linear correlations of anthropometric parameters and motor skills with overweight participants is statistically significant, is confirmed because of relevant degree of prediction.However, the second, separate hypothesis about statistical significance of linear correlation of anthropometric parameters and motor skills with normally nourished participants and the third separate hypothesis about statistical significance of linear connection of anthropometric parameters and motor skills with obese girls are not confirmed.
3. The attained results on the sample of overweight participants with 41% of explained mutual variance and two high linear correlations are defined with the significant value of p ≤ .05twostatistically significant pairs of canonical functions (roots) are interpreted as: (1) circular dimensionality of the skeleton and explosive strength and (2) subcutaneous fatty tissue and speed.
4. Extracted canonical factors indicate that the higher the values on the variables of predictive latent dimension circular dimensionality of the skeleton and body mass, the worse were the results on the variables of criterion latent dimension explosive strength, that is to say the girls who had higher values on variables of latent independent dimension of subcutaneous fatty tissue had worse results in variables dependent latent dimension speed.
The attained results will contribute to the more efficient individualization of teaching because diagnosing somatic-morphological-motor structures, orientation and selection, planning, programming, practice and work control will correspond to students' individual characteristics and skills.

Table 1 .
Descriptive statistic of morpho-motor variables

Table 2 .
Statistical significance of canonical factors in morpho-motoric space with overweight participants

Table 3 .
Statistical significance of the canonical factors in morpho-motor space with overweight participants

Table 4 .
The structure of canonical factors in the morphological set of variables with overweight participants