Introduction

The genetic diversity of a cultivated species represents the inheritable variation between and within its populations, and it is in this diversity that the selection processes applied by farmers and plant breeders operate. For this reason, it is necessary to characterize, take advantage of and preserve such genetic variability. In this respect, Mexico has a great diversity of maize (Zea mays L.), where each type is adapted to specific environmental conditions (^{Muñoz, 2005}).

^{Wellhausen et al. (1951)} mention that in Mexico the genetic variability of maize is the result of at least four important factors: the primitive breeds cultivated as free-pollinated varieties, the introduction of exotic varieties, natural crosses between teosinte and maize, and the country’s geography, which promotes differentiation between climatic types and agricultural systems.

In the American continent, between 220 and 300 maize races are considered (^{Brown and Goodman, 1977}); in Mexico maize has been cultivated for thousands of years, and perhaps in no other country is the range of genetic variation (^{Wellhausen et al., 1951}), since there are 41 (^{Ortega et al., 1991}), 59 (^{Sánchez et al., 2000}) and 65 races (^{LAMP, 1991}); however, the process of classification and identification is still incomplete.

One of the most important maize breeds in Mexico, because of its unique characteristics such as the length of its cob, is the Jala race. The earliest report states that at the beginning of the last century about 300 hectares of Jala maize, known locally as “wet maize” and whose ears measured up to 60 cm in length (^{Kempton 1924}; ^{Tibón and Beltrán, 1979}; ^{Listman and Pineda, 1992}). This breed has been cultivated exclusively in the Jala Valley, Nayarit, Mexico, for centuries (^{Aguilar et al., 2006}; ^{CIMMYT, 2007}), only now in an area of no more than 30 ha. The enormous length of the Jala ear is only expressed in the Jala Valley, which is why it is considered a unique ecotype of the Valley and of Mexico. This particularity allows to know the genetic and phenotypic changes of this race during short or long periods in a single geographical area.

The cob of the Jala maize is recognized worldwide as the largest (up to 60 cm in length), and its plant is high, with more than 5 m (^{Kempton, 1924}; ^{Rice et al., 2006}), and very late to harvest (seven or eight months) with poor adaptability (^{Valdivia et al., 2010}). ^{Muñoz (2003)} mentions that in the contest of the largest corn that is realized in Jala Nayarit, there were specimens with 71 cm in length. However, in the last 15 years this breed has suffered severe “genetic erosion” that has caused a dramatic reduction in cob and grain length (^{Aguilar et al., 2006}).

^{Aguilar et al. (2006)}; ^{Rice et al. (2006)} have focused efforts on recovering the original large ear type, but there is no known research in which this breed is formally used to develop specific improved maize to produce corn, for example, or for other uses. However, to exploit the genetic potential of the Jala maize breed, it could be combined with other types of maize with similar uses under the hypothesis that its genetic qualities may be recombined or transferred to other commonly used maize (^{Valdivia et al., 2010}).

Given the great importance of maintaining the genetic resources of Jala maize in its purest form, especially its large ear length characteristic, and with the purpose of proposing alternatives for its conservation and genetic improvement, as well as to evaluate its length on yield and other agronomic attributes, the present work evaluated native and semi-improved populations with germplasm of this breed in different environments of the states of Mexico and Nayarit.

Materials and methods

In this work, 14 maize populations representative of the Jala race or with germplasm of this breed were evaluated. Five populations were collected from farmers in the region of Jala, Nayarit, during 2012, who were named P1, P2, P3, P4 and P5. Seven populations were the individual selection compounds UAN-2008, UAN-2009A, UAN-2009B, UAN- 2009C, UAN-2010, UAN-2011, and Montecillo-2007 and two populations were the 13 XT and 8 XT varietal hybrids with 50% of Jala germplasm, which were used as controls for the yield of grain.

The populations were planted in 2012 in five localities (Table 1), three of which L1, L2, and L3 were established in plots of cooperating farmers in the State of Nayarit. The locality L4 was established in the facilities of the Academic Unit of Agriculture of the Autonomous University of Nayarit, Xalisco, Nayarit, and the L5 in the College of Postgraduates, Campus Montecillo, Texcoco, State of Mexico.

pH= pH del suelo.

Table 1 Soil and climatic characteristics and geographical location of the localities where the 14 maize populations were evaluated of Jala race. 2012. 

The establishment of the tests in the localities L1, L2, L3 and L4 coincided with the beginning of the rainy season; in the L5 was planted on May 10, where irrigation was applied until the normal rainy season was established. In each locality the evaluation was done in an experimental design of complete blocks at random with three repetitions. In all cases the experimental unit was two rows 5 m long and 0.8 m between rows, and two seeds were deposited per blow every 0.35 m. It should be noted that in these five experiments the same population density was used (70 000 plants ha^-1), but the farmer traditionally sowed at a lower density (from 30 000 to 40 000 plants ha^-1). Fertilization and cultivation were carried out in accordance with the traditional practices of each locality.

Results and discussion

The analysis of variance detected that between populations and between localities there was significant (p≤ 0.01) for all variables (Table 2), indicating that there were large phenotypic and genotypic differences between genotypes, even though they originated from the same genetic source which was the Jala race and the same geographic region, so that large genetic variation within the breed is inferred in addition, that the average productive potential of the environments was different, and that at least one of them was highly contrasting in its effects on genotypes with others. In the interaction Loc*Pop there was significance for nine variables (80%), indicating that at least one of the genotypes presented a different agronomic behavior for those characters when evaluated in different localities (Table 2).

FV= fuente de variación; FM= floración masculina; FF= floración femenina; AP= altura de planta; AM= altura de mazorca; ACT= acame de tallo; AMZ= aspecto de mazorca; LMZ= longitud de mazorca; DMZ= diámetro de mazorca; HMz= hileras por mazorca; GRh= granos por hilera; AGr= ancho de grano; LGr= longitud de grano; RGp= rendimiento de grano por planta; ^*, ^**, ns= diferencia significativa con α= 0.05, α= 0.01, no significativa, respectivamente; CV= coeficiente de variación; Pob= población; loc= localidad; rep (loc)= repeticiones anidado en localidades; Loc*Pob= interacción localidad por poblaciones; Rep= repetición.

Table 2 Mean squares of the combined variance analysis of 13 characters measured in 14 maize populations of the Jala race evaluated in five locations. 

The coefficients of variation of the combined analysis were generally low, ranging from 4 to 23%, except for the yield of cob per plant (RGp, 38%) and stemming (ACT, 26%) (Table 2). Which indicates the reliability of the information and that RGp and ACT were variables highly influenced by the environmental changes through localities, being highly quantitative.

With respect to the comparisons of means (Table 3), the variation between averages of the variables in general was wide. Thus, FM was represented by several significant groups, ranging from 77.2 to 89 dds, the FF ranged from 70 to 80.7 dds, the AP was 282 to 328.5 cm, the AM had values of 141.4 to 210.7 cm, the stalk resulted from 1.4 to 2.2, the cob aspect varied from 1.6 to 2.6, while the LMZ was 17.1 and 20.8 cm; the HMz were 11.6 to 13.4 rows; GRh ranged from 31.8 to 36.5 grains, AGr was 0.95 to 1.09 cm and LGr ranged from 1.7 to 1.9 cm. The cob yield per plant (RGp) showed no significant difference between averages, although values ranged from 82 to 131 g (5.7 to 9.1 t ha^-1) (Table 3). The DMZ did not show differences between averages, and its amplitude was of 4.6 to 4.9 cm. For the variable ear yield per plant, high significance was observed between genotypes (Table 2), and Tukey’s mean test detected significance only between the averages of P5 and Montecillo 2007 (Table 3), indicating that for the producer of maize 9.1 tonnes may be considered substantially and numerically more than 5.7 tonnes.

P1-P5= variedad de productor 1 a 5; FM= floración masculina; FF= floración femenina; AP= altura de planta; AM= altura de mazorca; ACT= acame de tallo; ASM= aspecto de mazorca; LMZ= longitud de mazorca; DMZ= diámetro de mazorca; HMz= hileras por mazorca; GRh= granos por hilera; AGr= ancho de grano; LGr=longitud de grano; RGp= rendimiento de mazorca por planta; DSM= diferencia significativa mínina^{. §}= medias con distinta letra en una columna son estadísticamente diferentes (Tukey, p≤ 0.05).

Table 3 Comparison of means (Tukey α= 0.05 probability) of 13 variables from 14 maize populations of the Jala race evaluated in five localities.  

With respect to the ear length (LMZ), the mean values of this in the five localities and the 14 maize populations evaluated are presented in Table 4. In the locality L1, the population that showed the highest expression in LMZ was UAN-2008, with 22 cm in L2 was producer 5 (P5) with an LMZ of 22.9. In the locality 3 Jala, the population that showed a greater length was UAN-2011, with an average of 20.8 cm, in the locality L4 the population that presented greater length was the one of the producer 4 (P4), with 19.9 cm, finally, in the locality L5, Montecillo, the population that had the greater expression was UAN-2011, with a length of 19.8 cm. These results evidenced that the maize of the maize populations of this work at present no longer present the previously reported lengths for the breed (^{Muñoz, 2005}), so genetic erosion and contamination is inferred in this, and possibly adverse effects of climate change, which has become more acute in recent years (^{Muñoz, 2014}).

P1-5= productor; L1= San José de Mojarras; L2= Ixtlán del Río; L3= Jala; L4= Xalisco; L5= Montecillos; β= fluctuación por localidad de la población que presentó mayor longitud de mazorca en las tres repeticiones. ^§= medias con distinta letra en una hilera son estadísticamente diferentes (Tukey, p≤ 0.05).

Table 4 Length of cob on average populations and environments. 

In general, it was observed that the UAN-2011 population was the one that presented a longer length, with an average of 20.8 cm, and the locality where it was expressed a greater length was the L2, that corresponded to Ixtlan of the Río, with an average of 20.7 cm cob (Table 4), a very distant value (71 cm) by ^{Muñoz (2003)}.

The locality L3 corresponded to Jala, Nayarit, and there the average length of cob was 17.7 cm, remaining in the last place through localities. It should be noted that this is the locality of origin of this breed, where in the past, ears of up to 60 cm or more in length have been documented (^{López, 2002}; ^{Rice et al., 2006}). The loss of the expression of ear length is due in large part to the tremendous shortage of seed of the typical Jala race, and also to the introduction of new crops, such as hybrid maize that contaminates the genetic purity of Jala, or sugarcane (Saccharum spp.), tobacco (Nicotina tabacum) and Jamaica (Hibiscus sabdariffa), which removed surface cultivation and diversity of the Jala race (^{Aguilar and Carballo, 2007}) and by the construction of new which also led to a reduction in planting area and loss of varieties.

It is worth mentioning that in the town of Jala a seed plot with characteristics similar to those traditionally used by farmers was sown, and where they have obtained the best expressions of ear and yield, but by atypical natural conditions (river overflow and stream), this plot was lost and it was only possible to count the data of the plot with the characteristics mentioned in Table 1. This could influence in the low expression of cob and low yield of grain obtained with the evaluated populations and sites.

However, with the results found in the evaluated populations of Jala maize, it can be mentioned that this breed has already suffered from severe “genetic erosion”, as according to ^{Rice et al. (2006)} in the recent past documented cobs of 60 cm or more in length, while in the present study the longest cob was found to be 33 cm, which is in agreement with ^{Wellhausen et al. (1951)}, who mention ear lengths of 30.5 cm. However, this contrasts with that reported by ^{Rice et al. (2006)}, in the sense that this breed has had a decrease in the length of the cob of 28 cm in the last 15 years; that is, 60 cm has been lowered to 32 cm on average.

As for cob yield per locality, the best numerical expression of this character was obtained in Xalisco and Ixtlan of Río, Nayarit, with 157 and 121 g (10.9 and 8.4 t ha^-1), respectively, and a significant difference between them of 36 g (2.5 t ha^-1). The high yield obtained in these localities could be attributed to a greater quantity and better distribution of the rainfall during the evaluation cycle, and the good characteristics of texture and soil fertility, opposite and contrasting in the locality of Jala, where its yield mean was 49 g (3.4 t ha^-1) (Table 5).

P1-5= productor; L1= San José de Mojarras; L2= Ixtlán del Río; L3= Jala; L4= Xalisco; L5= Montecillos. ^§= medias con distinta letra en una hilera son estadísticamente diferentes (Tukey, p≤ 0.05).

Table 5 Grain yield per plant (g) for populations and environments  

The previous result in Jala could be due to the presence of severe drought before flowering and during the period of grain filling, and to the non-representative production characteristics of the land used, such as sandy texture and low moisture retention and soil fertility , which resulted in this locality being the most unfavorable, with the lowest yield, even though it is recognized that Jala, Nayarit, is the site where only the largest ears of the Jala race of the world are obtained; that is to say, only there operates the ecotype under the favorable conditions.

The best populations yielded between 97 and 131 g (6.7 and 9.1 t ha^-1) per plant, and the worst ranged from 82 to 93 g (5.7 and 6.5 t ha^-1), with P5 being the best, (Montecillo, Texcoco, State of Mexico), where prevailing semidry temperate conditions of High Valley are very different from those of Valley of Jala, Nayarit (Table 1), and the Jala race, being an ecotype, does not flourish the same elsewhere outside the Valley, even with selection.

In the town of Xalisco, Nayarit, which had the best environmental conditions, populations with grain yields per plant were higher than 267 g (18.6 t ha^-1) (Table 5). The high yield obtained in the locality of Xalisco can be attributed to that its average of LMZ was of 20.2 cm, which was of the highest. In addition, P5 presented the highest LMZ, with an average of 20.6 cm, only 0.2 cm below the overall mean of 20.8 cm (Table 3).

In the Table 5 shows that in the locality of Jala, presented unfavorable conditions drought during flowering and stage of grain filling, the best population that was the hybrid control 8 XT, reached a maximum yield of 80 g (5.6 t ha^-1), which is acceptable, since it exceeded the national average of 3.5 t ha^-1 of grain (^{SIAP, 2015}) under adverse conditions. However, in the other localities some of the populations outperformed the control hybrids (Table 5), indicating their genetic potential as populations to be exploited in local and national breeding programs of the Jala race.

As conservation strategies and genetic improvement of the Jala maize race, based on the evidence obtained in this work, it is proposed that collections be made exhaustively throughout the Valley of Jala where it is grown, including all producers. This will make it possible to concentrate all of the germplasm currently available to make assessments of more populations and environments in the Jala Valley to find the one most promising as typical Jala maize. In this work, it was found that the P5 population was the maximum grain yield in the locality of Xalisco, Nayarit, and also had the highest ear length. This population could be used as a genetic basis to recover ear length and good grain yield in the Jala race through the recurrent selection practiced in environments of the Jala Valley in Nayarit on individuals typical of the Jala race.

Conclusions

According to the results, mean ear length varied between populations of 17.1 cm (8 XT) to 20.8 cm (UAN-2011), while for localities, Jala was the site with the lowest average, with a average of 17.7 cm, and Ixtlán del Río was the one that had the greatest expressions, with 20.7 cm. On the other hand, the highest yields by location and population were 157 g (10.9 t ha^-1) and 267 g (18.6 t ha^-1), which were obtained in Xalisco, Nayarit, and P5, respectively.

The P5 population had the cob length, with 22.9 cm, and a cob yield, with 131 g, which represents a good genetic basis to recover cob and good grain yield in the Jala breed through recurrent selection. On the other hand, it could be stated that the characteristic length of the Jala breed has been lost in the evaluated populations, since none of them had a general average greater than 30 cm, but in contrast, populations and localities with high yield capacity of grain, which is useful for the genetic improvement of the population base of the Jala breed by selection, both for cob size and yield.