Introduction

Ecuador has one of the highest biodiversity index values in the world, although its domestic animal populations are poorly studied. These are vital resources essential to the country’s food security and sovereignty¹. However, the most recent national reports on biodiversity² and agrobiodiversity³ state that deforestation, changes in land use, pollution and the introduction of exotic species are the main factors threatening agrobiodiversity. Incorporation of foreign livestock breeds is the main threat to conserving domestic food animal genetic resources.

The Domestic Animals Diversity Information System (DAD-IS) lists thirteen species of domestic food animals in Ecuador. Four of these are native Andean species (alpaca, 1; guinea pig, 1; llama, 1; vicuña, 1) and one is the native turkey (1). The remaining species are introduced: buffalo (1), cattle (21), goat (1), sheep (5), pig (8) and poultry (chicken, 1; and duck, 1). Cattle (Bos sp.) dominates Ecuadorian livestock production⁴. There are five European breed (Bos taurus) populations (Angus, Brown Swiss, Holstein, Jersey, and Normanda), and three Asian breed (Bos indicus) populations (Brahman, Gyr, Nelore). There are also ten creole-type populations (Bravo de Paramo, Chusco, Criollo Santa Elena Peninsula, Ecuadorian Creole, Esmeraldeño, Galapaqueño, Jaspeado Manabita, Macabea, Moro and Zarumeño), and three synthetic populations (Pizan, Sahiwal and Santa Gertrudis).

Research on the creole breeds of Latin America has found that the main problems reported by producers and technical advisors when using these breeds are lack of data and an absence of characterization and productive behavior studies⁵. This definitely holds true for the Criollo Santa Elena Peninsula (CSEP) cattle population. Cattle production is the main livestock activity in Santa Elena Province, Ecuador, and is mainly done using dual-purpose systems. Producers mainly use medium-sized herds.

The FAO⁶ considers it a priority to do breed characterization studies as the first phase in implementation of a livestock development program focused on a sustainability in traditional production systems that is linked to adequate land management. Characterization of animal genetic resources (AnGR) covers all activities associated with the identification, and quantitative and qualitative description of breed populations, and the natural habitat and production systems to which they are adapted⁷.

Descriptive biometric analysis has been widely used for breed characterization; for example, in a recent breed characterization of Criollo Manabí cattle in Ecuador⁸. Principal components analysis (PCA) is useful both in determining the relationship between biometric variables within a population⁹ and in differentiating between populations¹⁰. Discriminant analysis is normally used to analyze multivariate differences between groups, to determine the variables most useful in discriminating between groups, and to identify which groups are similar and which are different. It has recently been used in comparative morphometric studies of creole cattle breeds in Argentina¹⁰ and Africa¹¹, and of other domestic species: horses¹²; sheep¹³; goats¹⁴; pigs¹⁵; dogs¹⁶; ducks¹⁷; and turkeys¹⁸. Canonical discriminant analysis has been applied to productive characteristics in beef cattle^19,20 and in milking suitability²¹.

The present study objective study was to generate a biometric characterization of the Criollo Santa Elena Peninsula breed through a biometric analysis and morphometric differentiation of this breed compared to other Ecuadorian creole cattle breeds, with the goal of developing a purebred breeding program.

Material and methods

Results

Descriptive statistics of the morphometric variables and the ANOVA results using sex as the only variation factor showed that most of the variables exhibited moderate population variability (Table I). This largely confirms existence of discrete morphometric uniformity in the studied population, except for DSD, HL, IIW and FL in males, HW in females, and CF in both sexes. Variability was generally greater among the males. Most of the morphometric variables differed between males and females (P<0.001). The variables HW and RH differed moderately (P<0.05), and no difference was present for HL, CF, DSD, RH and IIW (P>0.05).

The descriptive statistics and ANOVA between the sexes for the CSEP morphometric indices showed all to have a generally moderate to high degree of variability in males, especially for the RTDI, ALD2, W5 and TORI (Table 2). The ALD1 and W6 indices were variable in both sexes, resulting in lower variability for the remaining indices for females. When compared between sexes almost all the indices differed (P<0.001), although the significance was lower in the RDIT (P<0.05), and no differences were apparent for PELI, ALD1 and W6 (P>0.05).

The Pearson correlation coefficients for the residuals of the analyzed variables (Table 3), showed a moderate degree of harmony in this population, 55.24% of the coefficients being significant (P<0.05). The correlations were high for the CL and TP variables, but less so for WH, RH, OIL, RL and IIW. The highest phenotypic correlation coefficient values were between BCD and CL, and RL and IIW (r = 0.86), and to a lesser extent between CBC and FL (0.75), WH and RH (0.71) and CL and FL (0.70).

The first six principal components explained 85 % of total variation (Table 4). Of all fourteen principal components (14), eight (57.0 %) had a value of less than 0.7. Four principal components (PC), accounted for 73.42 % of total variance (Table 5). Factor PC1 was identified with CL, which was characterized by negative correlations versus TP (-0.89), OIL (-0.82), WH (-0.75), RH (-0.75), RL (-0.74) and IWI (-0.70); that is, the animal’s body condition decreased as CL increased. This first factor explained 33.46 % of the variation in the original variables. Factor PC2 was associated with BCD, where an increase in this variable corresponded to increased rump size (RL, IIW) and lower CBC. This factor explained 21.82 % of the total variation. The next two factors are PC3, which was associated with CF and explained 10.92 % of variation, and PC4, which was linked to HW and explained 7.22 % of variation. The analysis of variance of the linear functions of the first six PC confirmed the greater weight of PC1 and PC2 by identifying significant differences between sexes only for PC1 and PC2, with the remaining components exhibiting statistical homogeneity.

Significance of the first two canonical discriminant functions produced was tested with Wilk’s Lambda (λ) (values of 0.03 and 0.22, respectively) and chi-square tests (χ²) (2,457.67 and 1,008.03, respectively; P≤0.001) (Table 6). Function 1 explained 72.33% of total variation and Function 2 explained 25.68 %; Function 3 explained less than 2% of the variance. These results validate the discriminant analysis, highlighting that Function 1 has the best linear combination of features that allow discrimination between the four studied populations.

A bidimensional graph of variables CAN 1 and CAN 2 shows the relationships between the four populations, with significant overlap between CM and CSTD. The CL and CSEP variables are clearly separate from the other variables, creating distinct groups with no overlap.

Figure 1 Graph of canonical discriminant analysis based on morphometric variables of females from four populations of Ecuadorian Creole cattle 

The Mahanalobis distances (upper diagonal) and Euclidean distances (lower diagonal) between the four populations clearly show the proximity between CM and CSDT (2.09) and the greater distance between CSEP and CSDT (47.56), considering that all values are significant (P<0.05) (Table 7). The individual Euclidean distances confirm the proximity between CM and CSDT and between CSEP and CL, while highlighting the distance between these two groups. Correct classification of individuals was 86.61 % for CM, 43.40 % for CSDT, 93.42 % for CSEP and 83.50 % for CL.

Discussion

Phenotypic variability among the morphometric variables for CSEP was higher than that reported in other Ecuadorian Creole cattle populations: Criollo Lojano²⁷; Criollo Macabeo²⁸; Criollo Manabita⁸; and Criollo of Santo Domingo de los Tsachilas²⁹. This was also the case when compared to other Latin American Creole breeds: Patagonian Creole in Argentina¹⁰; Criollo Saavedra in Bolivia³⁰; Criollo Pantaneiro in Brazil³¹; Criollo Casanare in Colombia³²; Barroso or Salmeco Creole in Guatemala³³; Chinampo³⁴ and Mixteco Creoles in Mexico³⁵; Pampas Chaco Creole in Paraguay³⁶; Criollo Limonero from Venezuela³⁷; and Uruguayan Creole³⁸, among others. It was also higher than in native breeds from Spain such as the red and black Berrenda breeds³⁹; the Serrana from Teruel⁴⁰; the black Andalusian⁴¹; the Pallaresa⁴²; and the Morucha⁴³.

The studied CSEP population was found to have an intermediate body format compared to other Ecuadorian Creole cattle breeds. It is larger than the Uruguayan Creole³⁸, Mixtec Creole³⁵ and Creole from Panama⁵, among others, but somewhat smaller than the Patagonian Creole¹⁰, Barroso or Salmeco Creole from Guatemala³³, and the Criollo Manabita⁸, among other populations. It was also smaller than the native Spanish breeds red and black Berrenda³⁹; Serrana de Teruel⁴⁰; Pallaresa⁴²; black Andalusian⁴¹; and Morucha⁴³. The intermediate format and size of CSEP are similar to those of Portuguese native cattle breeds⁴⁴. This population can therefore be characterized as having a typically normal body format, probably in response to the influence of Iberian breeds⁴⁵ and as an adaptive advantage in tropical environmental conditions⁸.

Among the morphometric indices, the studied population’s average CEFI value identifies it as a dolichocephalic type with HL predominating over HW. This coincides with index values for other Latin American Creole breeds such as the Criollo Saavedra in Bolivia³⁰; the Barroso or Salmeco Creole in Guatemala³³; Criollo Limonero from Venezuela³⁷; and Criollo Manabita in Ecuador⁸. It is also the case for native Spanish breeds such as the Asturian Valley, the Bruna of the Pyrenees, Parda de Montaña and Pirenaica⁴⁶, and the Serrana de Teruel⁴⁰. Of note is that dolichocephaly is much more pronounced in males than in females.

Average RTDI and DCI values are indicative of skeleton fineness and its association with milk production suitability, especially in females. These values were used to characterize predisposition to milking fitness within the studied CSEP populations. The OIL / WH value indicated that this population has a sublongilineal body proportion, another trait compatible with milking fitness and particularly notable in females. Like most native Spanish and Latin American Creole environmental breeds, the CSEP has a dorsolumbar line with an ascending caudal inclination, which favors movement in rough terrain.

Presence of moderate to high sexual dimorphism in the studied CSEP population based on the ANOVA for morphometric variables coincided with the profile of environmental type breeds with minimal selection⁴⁷. However, this dimorphism is not as pronounced as in the Uruguayan Creole³⁸, Criollo Macabeo²⁸ and Criollo Manabita breeds⁸. Diphorphism in the CSEP is also supported by the CEFI, TORI and RWI indices since they differ between the sexes (P<0.001). When considered in conjunction with the differences present between the remaining productive indices, this confirms the occurrence of low morphostructural uniformity between males and females in this population. This situation suggests that the CSEP may be undergoing a crossbreeding process through use of sires influenced by exotic breeds. Or, in what is essentially the same process, males and females in the CSEP are treated as two distinct subpopulations, receiving different genetic management. If this is the case it would explain the differences identified between the sexes for both the ethnological and productive type morphometric indices, as well as the greater intrinsic variability among males. In contrast, the PELI and ALD1 values were statistically homogeneous in both sexes. This coincides with the adaptive nature of the traits they represent since ascending caudal inclination of the dorsolumbar line is important to movement in difficult terrain (PELI), and pelvis width is linked to ease of parturition (ALD1).

The low level of correlation between the analyzed variables is indicative of the high underlying variability in this population. This is to be expected in this type of population, which historically has been heavily genetically manipulated by producers using mismatched criteria, without properly structured breeding programs, and in the complete absence of uniform breed trait selection criteria⁴⁷. Similar results have been reported for Criollo Manabí cattle in Ecuador⁸, as well as for the Spanish native Serrana de Teruel breed⁴⁰. This is another possible consequence of the use of sires influenced by improved exotic breeds.

The first principal component (PC1) explained more than one third of the observed variance. This component defined cephalic structure versus general animal morphostructure, such that the larger the CL the greater the reduction in their body format in terms of heights, lengths and diameters. The second principal component (PC2) was associated with body capacity, meaning that an increase in BCD improved rump morphostructure, which is linked to adequate pelvic canal width in females based on IIW and RL. Cranium length (CL) is therefore a defining variable in intuitive selection of animal morphostructure. In addition, BCD is clearly linked to body capacity and to rump size as an adaptive advantage for ease of parturition in females. Both variables must then be considered when establishing selection criteria in Ecuadorian Creole cattle.

The canonical discriminant analysis among females showed that each breed has a distinct morphometric pattern, implying clear morphometric differentiation between the four studied populations. This differentiation may be due to the reproductive isolation between them, as well as variation in body mass selection criteria, both of which are related to the geographical distance between these populations. These results are corroborated by the different Mahalanobis distances between the four populations, with CM and CSDT being nearest each other, CL in an intermediate position and CSEP the furthest from the rest.

Conclusions and implications

The CSEP has a medium body format with a normal tendency, sublongilinear body proportions and a dolichocephalic-type cranium. Skeletal structure in females is fine, indicating their suitability for milk production. Clear differences between the sexes in both the morphometric variables and indices confirm the presence of moderate to high sexual dimorphism in the studied population. Indeed, this suggests the coexistence of males and females as two subpopulations subject to different genetic management. The discriminant analysis effectively differentiated between the four analyzed Ecuadorian Creole cattle populations, confirming that Criollo Santa Elena Peninsula cattle are a distinct population with a specific morphometric pattern. The present study suggests that the Criollo Santa Elena Peninsula is a separate breed within the double-purpose tropical creole cattle breeds.