Scaling of Organ Weights in Macaca arctoides SUSAN G. LARSON Uniuersity of Wisconsin-Madison, Madison, Wisconsin 53706
KEY WORDS Allometry
. Scaling . Organ weights
ABSTRACT The allometric scaling of nine internal organs was examined for Macaca arctoides. Significant organ weight-body weight regressions were obtained for heart, lungs, kidneys, pancreas, thyroid, liver, and testes. The spleen and adrenal glands exhibited strong variability and were only loosely correlated to body weight. Using allometry as a criterion of subtraction, observed sex differences in mean organ weights were seen to be primarily the result of differences in average body weight. It is postulated t h a t analysis of observed differences in organ weights between this species and Macaca rnulatta would yield similar conclusions. Comparison of intraspecific slope values obtained in the present study with interspecific values reported in the literature reveals a pattern paralleling the brain-body weight relation. A discussion of the relationship between intra- and interspecific slopes is presented. Size variation in internal organs can be studied by a variety of techniques. In the past, many researchers have based their analyses upon mean organ weight values expressed as a percentage (or ratio) of body weight (e.g., Cameron, '25; Donaldson, '23; Howes et al., '60; Joseph, '08; Latimer, '39, '51, '65, '67). This practice reflects an underlying assumption that natural variability is represented by a linear relationship between organ weight and body weight. In other words, organ weight is thought to vary directly Le., geometrically) with body weight. However, Webster and coworkers (Webster et al., '47; Webster and Liljegren, '49, '55) have noted that organ weight ratios often change with increasing body size. Their findings indicate that the rates of growth of individual organs differ from the growth rate of the body as a whole. When rates differ in this manner, growth is said to be allometric. In more general terms, allometry can be defined as the study of such departures from geometric similarity that scale regularly with changes in size (Gould, '66). In most cases, the relationship between component body parts can be accurately described by an exponential function (the equation for simple allometry) y = bxa, where y is the organ in question and x is some measure of body size. AM. J. PHYS. ANTHROP. (1978)49; 95-102.
(Logarithmic transformation produces the most frequently used form of this equation i.e., the linear equation logy = log b a logx, where b = they-intercept and a = the slope.) Its simplicity and wide applicability (Huxley, '32; Brody, '45) have made this function an extremely useful tool in the analysis of size and shape in biology. Allometric analysis has been successfully applied to the study of organ weight variation in several species of laboratory animals (Addis and Gray, '5Oa,b,c; Christian, '67; Gray and Mahan, '43; Hall and MacGregor, '37; Mixner et al., '43; Walter and Addis, '391, however, comparatively little work has been among primate species. Organ weight variation in rhesus (Fremming et al., '55; Inay et al., '40; Kennard and Willner, '41; Kerr et al., '691,squirrel (Middleton and Rosal, '721, and howler (Stahl e t al., '68) monkeys has been studied in some detail, but primarily through nonallometric techniques. Although Stahl ('65) and Stahl and Gummerson ('67) are important exceptions, their investigations have been exclusively concerned with allometric organ weight scaling in primates as a group. No studies of intraspecific scaling for samples of adult primates have appeared in the literature to date, despite the necessity of such
+
95
96
SUSAN G. LARSON
work for understanding and comparing patterns of variation between species. Stahl (’65) had proposed t h e notion of a “basic mammalian physiological design” to which primates and other mammalian forms closely adhere. However, such broad generalizations, whatever their grounding in fact, should not deflect attention or research from smaller, species specific variation which may reflect differences in adaptation. Using allometry as a “criterion of subtraction” (Gould, ’66, ’75), intraspecific plots could be useful in separating those differences due to mechanical size requirements from those reflecting adaptive strategies not directly related to body size p e r se. This could apply equally to species differences as well as sex differences within species. Intraspecific analyses may also be useful in shedding light upon some “classical” problems of organ weight scaling. The best known organ-body size relationship-one which has been extensively studied both at the inter- and intraspecific levels-is t h a t of brain weight to body weight. Allometric analysis has revealed a consistent difference between t h e magnitudes of the slopes for inter- and intraspecific scaling. Although there have been several proposals, t h e factors underlying this difference remain largely unexplained (see Gould, ’75, for discussion). If other internal organs could be shown to follow a similar pattern, they would at least provide parallel cases for investigation. The present study treats t h e intraspecific scaling of organ weights in adult stumptail macaques f‘Macaca arctoidesl. The allometric equations with their statistical documentation are presented for each organ. In addition, mean values plus standard measures of variation are included to facilitate comparisons with similar data reported for other species. METHODS
Organ weights were obtained from autopsy records of t h e Wisconsin Regional Primate Research Center collected between 1965 and 1975. Each report contained a summary of t h e animal’s history, i t s physical condition at time of death, and detailed descriptions of t h e gross morphology and histology of each organ. A total of 170 adult stumptails had been examined during this 10-year period. Thirty of these were eliminated due to t h e absence of body weight data. The remaining 140 were carefully screened in order to collect a sample
TABLE 1
Causes of death among sample animals N
Cause of death
Veno-occlusive disease induced by monocrotaline intoxication a. Liver failure b. Sacrificed c. Anesthesia Induced renal dysfunction a. Renal failure b. Sacrificed c. Anesthesia Pneumonia Traumatic deaths Gastric disturbances Sacrificed --control animals Surgical shock Electrolyte imbalance
’
Total
32 10
6 6 1 1 7
6 4
3 2 2
80
Disease causing obliterating endophlebitls of small hepatic vein radicles leading to cirrhosis (Stedman’s Medical Dictionary. ‘721. An alkaloid In the seeds. leaves and stems of Crolalaria sprctnhrlis, a plant poisonous to liveatuck and poultry i n the southern IT S.(Stedman’s Medical Dictionary, ‘72)
of normal adults. An initial sample of 105 was assembled from which a final sample of 80 animals (39 males and 41 females) was ultimately obtained. Three major criteria were employed to eliminate unsuitable animals from t h e study sample: (1) use in a project that could affect normal growth and development, e.g., nutritional studies, experimental hypophysectomies, thyroidectomies, etc., or a n y treatments begun at a n early age; (2) general obesity or emaciation as well as experimental conditions t h a t could affect body weight; and (3) general debilitation for any reason, leading to the animal’s demise or euthanasia. Table 1contains a list of the causes of death among t h e sample animals. The majority of individuals were afflicted by some disease, but t h e detailed nature of t h e autopsy reports allowed differentiation of t h e unaffected organs for use in t h e analysis. Data was collected on t h e heart, liver, lungs, kidneys, pancreas, spleen, adrenals, thyroid, thymus, testes, and ovaries. However, weights for t h e thymus and ovaries were only infrequently recorded in t h e autopsy reports, hence these organs were not included in t h e analysis. Means, standard deviations, and coeffi-
’
Body (‘32) has commented on the difficulties In separating the thymus from surrounding connective and fatty tissues This and t h e small size of the ovaries probably contributed to their omission In t h e records Kegrettably, this resulted in sample sizes for these glands too small for meaningful analysis
97
SCALING OF ORGAN WEIGHTS
cients of variation were computed from the raw organ weights. These were included primarily to allow comparison with data for other species reported in the literature. In particular, data for t h e rhesus monkey reported by Kerr et al. ('69) were compared to the stumptail results using t-tests for significant mean organ weight differences. Similarly, t-tests were used to explore for significant sex differences within M.arctoides. The allometric analysis was carried out with the aid of a UNIVAC 1110 computer using the PICT 1 program in t h e Madison Academic Computing Center's STATJOB series. The program logarithmically transformed the organ weights (recorded in grams) and body weights (recorded in kilograms) and fitted them to regression lines by least squares. The computed allometric and statistical parameters include: t h e allometric or regression equation (logy = lob b a log XI, the standard error of regression (SER), the correlation coefficient (r),and a t-test for r. In addition to these, the following two parameters were calculated: (1) F-values to test the significance of the regression (i.e., whether a significant portion of t h e variance in the organ weight was explained by regression on body weight [Sokal and Rohlf, '7311, and (2) standard errors of the regression coefficients ha) for determination of confidence limits for a. Regression lines were fitted to each total sample and to males and females separately. Sex dif-
+
ferences in organ weight scaling were examined using t-tests for significant differences between male and female slopes and male and female y-intercepts. RESULTS
The results of the organ weight analysis are summarized in tables 2 and 3. In both tables the reported sample size for each organ is less than t h e total number of individuals ( 8 0 )used in the study. This is due to the elimination of diseased organs from analysis. Table 2 records means, standard deviations, percentages of body weight, and coefficients of variation for each of the organs studied. Males and females were found to differ significantly in overall body weight, and in mean heart, liver, thyroid, and pancreas weight (table 2). Comparison of the mean organ weights recorded here with those reported for the rhesus monkey by Kerr et al. ('69) revealed significant differences in mean kidney (p < 0.001), liver (p < 0.001), spleen (p < 0.011, adrenal (p < 0.051,and body weights (p < 0.01) between males, and in mean spleen weight (p < 0.01) between females. Table 3 includes values for the allometric parameters, standard errors of regressions (SER), correlation coefficients (r), F-tests for the significance of the regression, and results of tests for significant scaling differences between t h e sexes. Figure 1 includes two representative samples of organ weights regressed
TABLE 2
Mean organ weights for stumptail macaques Organ
Body wt
Heart Kidney
Sex
N
Mean
M F M F
39 41 36 40 28 27 31 30 17 16 29 34 10 15 25 31 28 32 29
7.91 k g 5.05 30.10 gm 21.99 34.52 30.93 12.16 8.96 56.27 46.15 1.24 0.96 235.04 169.69 12.39 10.52 2.32 2.14 36.24
M
F Pancreas
M F
Lung
M F
Thyroid
Liver Spleen Adrenal
M F M F M F
M F
Gonads
M
SD 2.99 1.44 12.38 5.92 5.77 7.90 4.81 3.10 16.73 16.14 0.62 0.37 73.41 49.45 3.93 4.38 0.84 0.63 14.34
1,B W
0.38 0.44 0.46 0.63 0.16 0.18 0.74 0.91 0.02 0.02 2.70 3.39 0.16 0.22 0.03 0.04 0.44
' Siwificance level for t ~ t e s t sof differences between means for males and females
cv 37.80 28.47 41.13 26.91 16.72 25.55 39.54 34.58 29.73 34.98 49.45 38.99 31.23 29.14 31.71 41.61 35.98 29.42 39.57
Probabilities
< 0.01 < 0.01 10.05
< 0.01 > 0.05 < 0.05 < 0.02
> 0.10 > 0.30
98
SUSAN C, LARSON
’c
:/
1.71
: 07
1.1
Ii
a
a
1-91
. a
a
;.-
a “
Y
a
a a
a
m
m 0
-I
.30-
.15.15
a
.525
a
a
a
*gl .375
a
.60
m
a
a
.00 .675
,825
.975
1.125
1.275
a *
a I
.30
I
I
.45
1
f
.60
L o g BW
a
.90
.75
1
1
1.05
1.20
L o g BW
A
B
Fig. 1 Two representative samples of organ weight regressed against body weight after logarithmic transformation. (A) Narrow point scatter of organ closely correlated to body weight. (B) Wide point scatter of organ only loosely correlated to body weight. HW, heart weight; AW, adrenal weight; BW, body weight.
TABLE 3
Allometric parameters for organ weight scaling in stumptail macaques Organ
Heart Kidney Pancreas Lung Thyroid Liver Spleen Adrenal Gonads
Sex
N
M 36 F 40 M + F 76 28 M F 27 M + F 55 M 31 30 F M + F 61 17 M 16 F M + F 33 M 29 F 34 M + F 63 M 10 15 F M + F 25 M 25 F 31 M + F 56 M 28 F 32 M + F 60 29 M
Lo&
0.82 0.83 0.84 1.28 1.15 1.24 0.58 0.46 0.50 1.28 1.09 1.22 -0.34 -0.54 -0.44 2.13 1.79 1.87 0.76 0.97 0.81 0.07 0.12 0.12 0.22
u
0.72 0.73 0.70 0.30 0.48 0.35 0.56 0.67 0.63 0.53 0.81 0.61 0.47 0.70 0.56 0.24 0.62 0.51 0.35 0.02 0.29 0.32 0.28 0.27 1.43
95X CL,
0.83-0.57 0.48-0.23
r
Prob(r)
0.76 0.70 0.78 0.72 0.53 0.62 0.60
< 0.001
0.50
0.83-0.43 0.85-0.37 0.82-0.29 0.82-0.20 0.54-0.03 0.48-0.07 1.80-1.07
0.64 0.58 0.69 0.68 0.38 0.48 0.47 0.23 0.52 0.58 0.38 0.01 0.29 0.37 0.25 0.33 0.84
47.47 36.97 0.50
> 0.50
> 0.30
> 0.50
>0.50
>0.50
0.50
>0.40
< 0.001 0.10
KEY WORDS Allometry
. Scaling . Organ weights
ABSTRACT The allometric scaling of nine internal organs was examined for Macaca arctoides. Significant organ weight-body weight regressions were obtained for heart, lungs, kidneys, pancreas, thyroid, liver, and testes. The spleen and adrenal glands exhibited strong variability and were only loosely correlated to body weight. Using allometry as a criterion of subtraction, observed sex differences in mean organ weights were seen to be primarily the result of differences in average body weight. It is postulated t h a t analysis of observed differences in organ weights between this species and Macaca rnulatta would yield similar conclusions. Comparison of intraspecific slope values obtained in the present study with interspecific values reported in the literature reveals a pattern paralleling the brain-body weight relation. A discussion of the relationship between intra- and interspecific slopes is presented. Size variation in internal organs can be studied by a variety of techniques. In the past, many researchers have based their analyses upon mean organ weight values expressed as a percentage (or ratio) of body weight (e.g., Cameron, '25; Donaldson, '23; Howes et al., '60; Joseph, '08; Latimer, '39, '51, '65, '67). This practice reflects an underlying assumption that natural variability is represented by a linear relationship between organ weight and body weight. In other words, organ weight is thought to vary directly Le., geometrically) with body weight. However, Webster and coworkers (Webster et al., '47; Webster and Liljegren, '49, '55) have noted that organ weight ratios often change with increasing body size. Their findings indicate that the rates of growth of individual organs differ from the growth rate of the body as a whole. When rates differ in this manner, growth is said to be allometric. In more general terms, allometry can be defined as the study of such departures from geometric similarity that scale regularly with changes in size (Gould, '66). In most cases, the relationship between component body parts can be accurately described by an exponential function (the equation for simple allometry) y = bxa, where y is the organ in question and x is some measure of body size. AM. J. PHYS. ANTHROP. (1978)49; 95-102.
(Logarithmic transformation produces the most frequently used form of this equation i.e., the linear equation logy = log b a logx, where b = they-intercept and a = the slope.) Its simplicity and wide applicability (Huxley, '32; Brody, '45) have made this function an extremely useful tool in the analysis of size and shape in biology. Allometric analysis has been successfully applied to the study of organ weight variation in several species of laboratory animals (Addis and Gray, '5Oa,b,c; Christian, '67; Gray and Mahan, '43; Hall and MacGregor, '37; Mixner et al., '43; Walter and Addis, '391, however, comparatively little work has been among primate species. Organ weight variation in rhesus (Fremming et al., '55; Inay et al., '40; Kennard and Willner, '41; Kerr et al., '691,squirrel (Middleton and Rosal, '721, and howler (Stahl e t al., '68) monkeys has been studied in some detail, but primarily through nonallometric techniques. Although Stahl ('65) and Stahl and Gummerson ('67) are important exceptions, their investigations have been exclusively concerned with allometric organ weight scaling in primates as a group. No studies of intraspecific scaling for samples of adult primates have appeared in the literature to date, despite the necessity of such
+
95
96
SUSAN G. LARSON
work for understanding and comparing patterns of variation between species. Stahl (’65) had proposed t h e notion of a “basic mammalian physiological design” to which primates and other mammalian forms closely adhere. However, such broad generalizations, whatever their grounding in fact, should not deflect attention or research from smaller, species specific variation which may reflect differences in adaptation. Using allometry as a “criterion of subtraction” (Gould, ’66, ’75), intraspecific plots could be useful in separating those differences due to mechanical size requirements from those reflecting adaptive strategies not directly related to body size p e r se. This could apply equally to species differences as well as sex differences within species. Intraspecific analyses may also be useful in shedding light upon some “classical” problems of organ weight scaling. The best known organ-body size relationship-one which has been extensively studied both at the inter- and intraspecific levels-is t h a t of brain weight to body weight. Allometric analysis has revealed a consistent difference between t h e magnitudes of the slopes for inter- and intraspecific scaling. Although there have been several proposals, t h e factors underlying this difference remain largely unexplained (see Gould, ’75, for discussion). If other internal organs could be shown to follow a similar pattern, they would at least provide parallel cases for investigation. The present study treats t h e intraspecific scaling of organ weights in adult stumptail macaques f‘Macaca arctoidesl. The allometric equations with their statistical documentation are presented for each organ. In addition, mean values plus standard measures of variation are included to facilitate comparisons with similar data reported for other species. METHODS
Organ weights were obtained from autopsy records of t h e Wisconsin Regional Primate Research Center collected between 1965 and 1975. Each report contained a summary of t h e animal’s history, i t s physical condition at time of death, and detailed descriptions of t h e gross morphology and histology of each organ. A total of 170 adult stumptails had been examined during this 10-year period. Thirty of these were eliminated due to t h e absence of body weight data. The remaining 140 were carefully screened in order to collect a sample
TABLE 1
Causes of death among sample animals N
Cause of death
Veno-occlusive disease induced by monocrotaline intoxication a. Liver failure b. Sacrificed c. Anesthesia Induced renal dysfunction a. Renal failure b. Sacrificed c. Anesthesia Pneumonia Traumatic deaths Gastric disturbances Sacrificed --control animals Surgical shock Electrolyte imbalance
’
Total
32 10
6 6 1 1 7
6 4
3 2 2
80
Disease causing obliterating endophlebitls of small hepatic vein radicles leading to cirrhosis (Stedman’s Medical Dictionary. ‘721. An alkaloid In the seeds. leaves and stems of Crolalaria sprctnhrlis, a plant poisonous to liveatuck and poultry i n the southern IT S.(Stedman’s Medical Dictionary, ‘72)
of normal adults. An initial sample of 105 was assembled from which a final sample of 80 animals (39 males and 41 females) was ultimately obtained. Three major criteria were employed to eliminate unsuitable animals from t h e study sample: (1) use in a project that could affect normal growth and development, e.g., nutritional studies, experimental hypophysectomies, thyroidectomies, etc., or a n y treatments begun at a n early age; (2) general obesity or emaciation as well as experimental conditions t h a t could affect body weight; and (3) general debilitation for any reason, leading to the animal’s demise or euthanasia. Table 1contains a list of the causes of death among t h e sample animals. The majority of individuals were afflicted by some disease, but t h e detailed nature of t h e autopsy reports allowed differentiation of t h e unaffected organs for use in t h e analysis. Data was collected on t h e heart, liver, lungs, kidneys, pancreas, spleen, adrenals, thyroid, thymus, testes, and ovaries. However, weights for t h e thymus and ovaries were only infrequently recorded in t h e autopsy reports, hence these organs were not included in t h e analysis. Means, standard deviations, and coeffi-
’
Body (‘32) has commented on the difficulties In separating the thymus from surrounding connective and fatty tissues This and t h e small size of the ovaries probably contributed to their omission In t h e records Kegrettably, this resulted in sample sizes for these glands too small for meaningful analysis
97
SCALING OF ORGAN WEIGHTS
cients of variation were computed from the raw organ weights. These were included primarily to allow comparison with data for other species reported in the literature. In particular, data for t h e rhesus monkey reported by Kerr et al. ('69) were compared to the stumptail results using t-tests for significant mean organ weight differences. Similarly, t-tests were used to explore for significant sex differences within M.arctoides. The allometric analysis was carried out with the aid of a UNIVAC 1110 computer using the PICT 1 program in t h e Madison Academic Computing Center's STATJOB series. The program logarithmically transformed the organ weights (recorded in grams) and body weights (recorded in kilograms) and fitted them to regression lines by least squares. The computed allometric and statistical parameters include: t h e allometric or regression equation (logy = lob b a log XI, the standard error of regression (SER), the correlation coefficient (r),and a t-test for r. In addition to these, the following two parameters were calculated: (1) F-values to test the significance of the regression (i.e., whether a significant portion of t h e variance in the organ weight was explained by regression on body weight [Sokal and Rohlf, '7311, and (2) standard errors of the regression coefficients ha) for determination of confidence limits for a. Regression lines were fitted to each total sample and to males and females separately. Sex dif-
+
ferences in organ weight scaling were examined using t-tests for significant differences between male and female slopes and male and female y-intercepts. RESULTS
The results of the organ weight analysis are summarized in tables 2 and 3. In both tables the reported sample size for each organ is less than t h e total number of individuals ( 8 0 )used in the study. This is due to the elimination of diseased organs from analysis. Table 2 records means, standard deviations, percentages of body weight, and coefficients of variation for each of the organs studied. Males and females were found to differ significantly in overall body weight, and in mean heart, liver, thyroid, and pancreas weight (table 2). Comparison of the mean organ weights recorded here with those reported for the rhesus monkey by Kerr et al. ('69) revealed significant differences in mean kidney (p < 0.001), liver (p < 0.001), spleen (p < 0.011, adrenal (p < 0.051,and body weights (p < 0.01) between males, and in mean spleen weight (p < 0.01) between females. Table 3 includes values for the allometric parameters, standard errors of regressions (SER), correlation coefficients (r), F-tests for the significance of the regression, and results of tests for significant scaling differences between t h e sexes. Figure 1 includes two representative samples of organ weights regressed
TABLE 2
Mean organ weights for stumptail macaques Organ
Body wt
Heart Kidney
Sex
N
Mean
M F M F
39 41 36 40 28 27 31 30 17 16 29 34 10 15 25 31 28 32 29
7.91 k g 5.05 30.10 gm 21.99 34.52 30.93 12.16 8.96 56.27 46.15 1.24 0.96 235.04 169.69 12.39 10.52 2.32 2.14 36.24
M
F Pancreas
M F
Lung
M F
Thyroid
Liver Spleen Adrenal
M F M F M F
M F
Gonads
M
SD 2.99 1.44 12.38 5.92 5.77 7.90 4.81 3.10 16.73 16.14 0.62 0.37 73.41 49.45 3.93 4.38 0.84 0.63 14.34
1,B W
0.38 0.44 0.46 0.63 0.16 0.18 0.74 0.91 0.02 0.02 2.70 3.39 0.16 0.22 0.03 0.04 0.44
' Siwificance level for t ~ t e s t sof differences between means for males and females
cv 37.80 28.47 41.13 26.91 16.72 25.55 39.54 34.58 29.73 34.98 49.45 38.99 31.23 29.14 31.71 41.61 35.98 29.42 39.57
Probabilities
< 0.01 < 0.01 10.05
< 0.01 > 0.05 < 0.05 < 0.02
> 0.10 > 0.30
98
SUSAN C, LARSON
’c
:/
1.71
: 07
1.1
Ii
a
a
1-91
. a
a
;.-
a “
Y
a
a a
a
m
m 0
-I
.30-
.15.15
a
.525
a
a
a
*gl .375
a
.60
m
a
a
.00 .675
,825
.975
1.125
1.275
a *
a I
.30
I
I
.45
1
f
.60
L o g BW
a
.90
.75
1
1
1.05
1.20
L o g BW
A
B
Fig. 1 Two representative samples of organ weight regressed against body weight after logarithmic transformation. (A) Narrow point scatter of organ closely correlated to body weight. (B) Wide point scatter of organ only loosely correlated to body weight. HW, heart weight; AW, adrenal weight; BW, body weight.
TABLE 3
Allometric parameters for organ weight scaling in stumptail macaques Organ
Heart Kidney Pancreas Lung Thyroid Liver Spleen Adrenal Gonads
Sex
N
M 36 F 40 M + F 76 28 M F 27 M + F 55 M 31 30 F M + F 61 17 M 16 F M + F 33 M 29 F 34 M + F 63 M 10 15 F M + F 25 M 25 F 31 M + F 56 M 28 F 32 M + F 60 29 M
Lo&
0.82 0.83 0.84 1.28 1.15 1.24 0.58 0.46 0.50 1.28 1.09 1.22 -0.34 -0.54 -0.44 2.13 1.79 1.87 0.76 0.97 0.81 0.07 0.12 0.12 0.22
u
0.72 0.73 0.70 0.30 0.48 0.35 0.56 0.67 0.63 0.53 0.81 0.61 0.47 0.70 0.56 0.24 0.62 0.51 0.35 0.02 0.29 0.32 0.28 0.27 1.43
95X CL,
0.83-0.57 0.48-0.23
r
Prob(r)
0.76 0.70 0.78 0.72 0.53 0.62 0.60
< 0.001
0.50
0.83-0.43 0.85-0.37 0.82-0.29 0.82-0.20 0.54-0.03 0.48-0.07 1.80-1.07
0.64 0.58 0.69 0.68 0.38 0.48 0.47 0.23 0.52 0.58 0.38 0.01 0.29 0.37 0.25 0.33 0.84
47.47 36.97 0.50
> 0.50
> 0.30
> 0.50
>0.50
>0.50
0.50
>0.40
< 0.001 0.10
