Leaf Architecture in the Morphological Diversity of the Genus Prosopis in the Semi-Desert Area of Northeastern Mexico

Abstract

Mesquite (Prosopis spp.) is one of the main plant representatives in regions with a dry climate, and is a fundamental part of the flora of the Mexican arid, with an indisputable importance from ecological, economic and industrial points of view. However, the restrictive factors of dry climates, as well as genetic variability, are sources of the great diversity of mesquite, so its taxonomy is not yet well defined. The present study seeks to determine the diversity of mesquite in the Mexican semi-desert based on the morphometric characterization of its leaves. Methods: Different leaf parameters such as the number, length and width of the leaflets were recorded in 31 well-marked sites in the area, to obtain measures of central tendency and dispersion, and to determine the differences and similarities between the sites, as well as the groups of homogeneous and heterogeneous taxa. Results: Five taxa were identified, of which there were two varieties (Prosopis glandulosa var. glandulosa and Prosopis reptans var. cinerascens), a pure or typical species (Prosopis laevigata) and two hybrids (Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa). The discriminant analysis indicated that five variables presented the highest percentage of separation or best separated the taxa, so the study was based on the phenogram with the combination of these five variables: (a) length of middle leaflets, (b) length of upper leaflets, (c) spacing of middle leaflets, (d) length/width relationship of middle leaflets and (e) length/width relationship of middle leaflets. Conclusions: Both the traditional taxonomic classification and the cluster and discriminant analyses reflected the same taxa (Prosopis glandulosa var. glandulosa, Prosopis laevigata and Prosopis reptans var. cinerascens) and the existence of hybridization between Prosopis glandulosa var. glandulosa and Prosopis laevigata.

1. Introduction

Prosopis is a genus that is distributed in all arid and semi-arid areas of the world, constituting one of the main plant resources in these regions. The species of this genus are presented as a transcendental resource for human life, since their use has continued over time by indigenous people, colonizers, pioneers and current rural residents. That is why it is considered one of the most useful plant resources in arid areas, since firewood, charcoal, chips, rubber, construction materials, doors, furniture, floors, fence posts, food, forage, nectar for bee-keeping, shade, work tools and medicine have been obtained from it [1]. On the other hand, mesquite plays an important role in the environment as a nitrogen-fixing plant, enriching the soil around it and promoting the growth of shrubs associated with it, therefore preventing soil erosion; it also acts as a nurse plant for numerous species of birds and rodents [2]. A good number of herbivorous insects feed on mesquite leaves [3]. The abundant flowers, which produce a copious amount of nectar and pollen, are exploited by an extensive group of insects, particularly bees [4]. Mesquite provides food and protection to many types of wild animals (birds and mammals), so it plays a valuable role in the conservation of wildlife and surely represents an important factor that helps maintain the complexity of ecosystems in arid and semi-arid zones [5].

Despite all these uses that demonstrate its important ecological, economic and industrial role, there is much confusion in the genus Prosopis, since its taxonomy is subject to almost continuous revisions, especially due to its amphitropical character. These taxonomic confusions within the genus were largely settled with the authoritative monograph of Burkart [6], who defined the generic limits, dividing the genus Prosopis into five sections, according to their floral characteristics, with each section presenting well-marked vegetative differences. The section Anonychium contains a solitary and unarmed species. The section Prosopis contains three species that all have internodal prickles, similar to those found on roses (Rosa spp.). The section Strombocarpa contains nine species, with spiny stipules, and the section Algarobia contains thirty species, with cauline, mostly axillary thorns. Between these latter two sections is the section Monilicarpa, with a single member. The section Algarobia has been the subject of numerous morphological, molecular and cytological studies. Burkart [6] identified five series in this section, separated on vegetative grounds, but not all have been confirmed by more recent work. Graham [7] demonstrates the existence of a narrow contact zone or zone of morphic intergradation between the long-leaved Prosopis glandulosa and “short-leaf” mesquite (included in P. laevigata by Johnston [8]). M.K. Johnston [8], following Graham [7] and Benson [9], drew attention to the importance of quantitative characteristics of foliage, as well as their geographical segregation and taxonomic value for determining the boundaries of species in the Prosopis section Algarobia. He also mentioned the taxonomic value of mature fruits, which, however, are not always present on living plants and in herbarium specimens. Johnston [8] believed that the characters of the inflorescences of Prosopis section Algarobia may have a certain taxonomic and diagnostic value, which, however, is difficult to assess.

On a molecular level, Bessega et al. [10] studied the internal transcribed spacers ITS1-5.8S-ITS2 sequences in many Prosopis species and noted weak genetic differentiation within the clade Algarobia–Monilicarpa, concluding that this result, together with the relatively low bootstrap supports of many clades within this group, may indicate recent divergence among the species within the clade. This highlights a lack of congruence between molecular and morphological data in the group Algarobia–Monilicarpa, so Bessega et al. [10] proposed several hypotheses to explain this phenomenon, among them the rapid parallel evolution of morphological traits in response to environmental factors, ancestral polymorphism of molecular markers and past introgressive events. On the other hand, Castillo et al. [11] studied deoxyribonucleic acid (DNA) sequences of the nuclear external transcribed spacer region (ETS) and two chloroplast intergenic spacers (rpl32-trnL and psbA-trnH) and revealed extremely low sequence variability across the initial subset of Prosopis species, including members of Algarobia section, from native and non-native populations. They concluded that these results suggest a recent radiation of these species and possibly incomplete reproductive isolation between them, which may lead to frequent hybridization and introgression between species in this section. Cytological studies revealed that the Prosopis species are most likely diploids with 2n = 28 [12,13]; so, if interspecific hybridization exists, it occurs on homoploid level.

From the above, it appears that the Prosopis taxonomy is very difficult due to the extreme variability between individual plants, even between individuals within a well-defined species. The nature of the morphology of most species varies with site and possibly also with inherent genetic factors [14]. In this way, the taxonomic classification of this group presents complications to be defined correctly, making it difficult to understand its distribution and evolution. On the other hand, the identification of these plants is still very difficult because they present great polymorphism due to environmental conditions and natural crossings between populations, which are facilitated by their enormous genetic plasticity [15].

In arid and semi-arid climates, there are several stress factors that limit the development and survival of plants: the availability of water and temperatures restrict their growth to the spring months, while the lack of precipitation, high temperatures and water stress of the summer months produces the cessation of growth processes [16]. Plants that can survive, develop and reproduce in these environments (arid and semi-arid) characterized by a variable period of water deficit, present morphological and physiological adaptations that allow them to evade or tolerate the limiting conditions of the environment, forming the so-called xeromorphic species [17]. Leaves are one of the organs with the greatest plasticity in plants, as they can modify their size and structural characteristics. Variations in the anatomy of the photosynthetic tissue, such as its thickness, pubescence and the relationship between the surface of the cell walls of the chlorenchyma and the outer surface, serve as an adaptation of the plant to hot and dry environments [18].

Based on the above, this work was carried out with the objective of differentiating mesquite species from different ecotypes of the semi-arid area of Mexico, through a cluster analysis in leaf morphometry, characterizing the variability, diversity and leaf morphometric parameters of populations through discriminant analysis.

2. Materials and Methods

2.1. Description of the Study Area

The study area is in the northeast of the Mexican Republic, with a precise geographical location between the parallels 23°10′27″ and 27°46′06″ north latitude, and the meridians 98°26′24″ and 101°13′55′ west longitude. The average annual temperature is 14 °C in the Sierra Madre Oriental, varying up to 24 °C in the Plains of North America, with 13 different climates, as shown in

Table 1. Types of climates identified in the northeast of Mexico and their symbology.

Type or Subtype	Weather Symbol	% on Surface
Very warm and warm semi-dry	BS1(h)	17.92
Very warm and warm dry	BS(h′)	17.15
Semi-warm dry	BSh	13.27
Semi-warm semi-humid with little rain all year round	ACx	10.44
Semi-warm semi-humid with rain in summer	ACw	8.66
Temperate dry	BSk	7.87
Temperate semi-dry	BS1k	6.65
Semi-warm semi-dry	BS1h	5.77
Semi-warm very dry	BWh	4.96
Temperate semi-humid with rains in summer	C(w)	4.52
Temperate semi-humid with little rain all year round	Cx	2.52
Semi-cold semi- humid with rains in summer	C(E)(w)	0.24
Semi-cold semi-humid with little rain all year round	C(E)(x)	0.03

Source. INEGI [19].

.

2.2. Description of Mesquite Species

The descriptive characteristics of the species of the genus Prosopis in the arid zones of Mexico are shown in

Table 2. Botanical characteristics, synonymy and distribution of Prosopis spp. in northeastern Mexico.

Species	Distribution *	Characteristics	Synonymy
P. glandulosa	Mina [20] Linares and Dr. Arroyo [21]	Oblong leaflets, less than five times longer than widths or lengths of 15 to 20 mm. Right sheath with recurved appendage, sometimes curved, yellow to brown [14].	Prosopis juliflora (Swartz) DC P. juliflora var. glandulosa Cockerell P. juliflora var. torreyana L. Benson [22]
P. glandulosa var. torreyana	Iturbide [20]	Leaflets less than five times longer than wide, herbaceous, relatively tender. They are usually trees. Distribution: coastal areas of Mexico [14].	Prosopis juliflora (Sw.) DC. var. torreyana L.D. Benson Prosopis odorata Torr. & Frém. [22]
P. glandulosa var. glandulosa	Mina and China [20]	Evergreen tree 1–10(–20) m high; one pair of pinnae per leaf: six to fifteen pairs of leaflets per pinna, obovate, oblong to narrow, glabrous, 30–40 mm long, 10–30 mm wide [14].	Algarobia glandulosa (Torrey) Cooper. Neltuma constricta (Sarg.) Britton & Rose Prosopis chilensis (Molina) Stuntz variety glandulosa (Torrey) Standley Prosopis juliflora (Sw.) DC. var. constricta Sarg. [23]
P. laevigata	Mina, Linares, China and Dr. Arroyo [20], Los Ramones [23]	Small leaves, shorter than or equal to the length of the inflorescence. Dry, thin, flexible and soft pod, compressed between the segments and whitish in color. Leaves with one or two pairs of pinnae and six to twenty pairs of leaflets. Distribution: northwest Mexico and the Baja California peninsula [14].	Acacia laevigata Willd. Algorobia dulces Benth Mimosa laevigata (Wild.) Poiret Mimosa rotundata Sessé & Mocino Nelthuma laevigata (Wild.) Britton & Rose Neltuma attenuata Prosopis dulcis Kunth [23]
P.reptans var. cinerascens	Bustamante, Villaldama [24].	Smooth pods with numerous compact regular whorls, united singly or only a few in a cluster [14].	Prosopis cinerascens (A. Gray) Benth. Strombocarpa cinerascens A. Gray [22]

Source: Guerrero [20]. * The distribution of mesquite is based on that reported by the authors’ studies.

.

2.3. Sampling Design

Based on the soil and vegetation use map with a scale of 2/50,000, 31 localities were selected, and simple random sampling was applied to collect leaves from Mezquite trees (Prosopis spp.). The design consisted of the random selection of 10 adult Mezquite trees that represent a “locality” or “site.” For each tree, 10 mature leaves were cut and collected in different vertical strata, to take measurements of the morphometric parameters. In addition, the ArcView version 3.2 program was used to geographically position the 31 sites (Figure 1 and

Table 3. Geographic location of the 31 sites sampled in the northeastern Mexico semi-desert area.

SAMPLING SITES			UTM COORDINATES
	LOCALITY	MUNICIPALITY	East	North	Elevation
1	Plan del Orégano	Melchor Ocampo	45°76′94″	28°83′106″	152
2	Colorados de Arriba	Vallecillo	40°09′06″	29°27′506″	239
3	El Llano	Los Ramones	43°70′24″	28°54′589″	193
4	La Barretosa	Los Herreras	45°19′90″	28°54′589″	166
5	Ej. Emiliano Zapata	Parás	43°16′51″	29°37′498″	164
6	Ej. Puente del Río Salado	Anáhuac	41°34′60″	29°82′484″	146
7	Loma Larga	General Treviño	44°88′58″	29°04′109″	147
8	Los Ébanos	Los Ramones	45°35′17″	28°24′931″	196
9	Ejido El Álamo	Vallecillo	42°10′68″	29°29′095″	195
10	Dulces Nombres	Pesquería	39°42′36″	28°44′953″	351
11	Hacienda San Pedro	General Zuazua	38°40′66″	28°67′217″	369
12	Higueras	Higueras	39°81′19″	28°72′672″	503
13	Los Pajaritos	Doctor González	40°88′54″	28°63′751″	423
14	Rancho el Recuerdo	General Terán	42°76′56″	28°06′597″	285
15	Loma la Parada	Marín	40°18’47″	28°56’144″	323
16	El Bajío	Marín	40°21′22″	28°58′136″	342
17	El Resumidero	Salinas Victoria	37°37′12″	28°82′558″	451
18	El Puente	Salinas Victoria	37°22′52″	28°72′246″	424
19	Rancho Gomas	Salinas Victoria	35°31′64″	28°95′485″	191
20	Km80 **	Los Ramones	44°60′57″	28°37′409″	191
21	Rancho La bonanza	General Terán	46°71′87″	27°87′506″	217
22	Rancho Nuevo	General Terán	45°28′41″	27°90′393″	259
23	Ejido las Presas	Lampazos	34°84′05″	29°82′066″	341
24	El Nogal	Anáhuac	39°91′25″	30°05′741″	177
25	Regantes 26	Anáhuac	36°71′40″	30°24′331″	228
26	Rancho la Ceja	Los Aldama	47°69′56″	28°83′426″	112
27	China	China	47°44′39″	28°45′014″	138
28	San pedro de los Escobedos	Linares	46°23′56″	27°60′631″	261
29	San Ignacio de Texas	Galeana	37°90′92″	26°90′496″	1684
30	Ejido las Puentes	Aramberri	39°06′06″	26°70′446″	1581
31	Km80 **	Los Ramones	44°91′25″	28°50′504″	188

** Sites 20 and 31 were labeled with the same locality name, but presented different species of Prosopis spp.

).

2.4. Study of Leaf Parameters

The trees sampled were mature individuals with the characteristic physiognomy of the species. The sampled leaves were taken from the middle part of the shoots to ensure that they were mature leaves with complete development. The morphometric parameters of the leaves (Figure 2) were measured. A digital vernier from the “Mitutoyo” brand was used to in trees from the 31 sites, recording the following 14 variables and three determinations corresponding to the existing relationships between the variables*: length of the petiole (mm), number of pinnae, width of pinna (mm), number of pairs of leaflets per pinna, length of rachis (mm), length and width of lower leaflet (mm), length and width of middle leaflet (mm), length and width of the upper leaflet (mm) and the following spacing between the leaflets: (A) lower: spacing between the leaflet immediately above the first pair of lower leaflets of the leaf (mm), (B) middle leaflets: spacing immediately below the middle leaflets of the leaf (mm) and (C) upper leaflets: immediate lower spacing of the last upper leaflets of the leaf (mm). *Length/width ratio for the lower, middle and upper leaflets.

2.5. Statistical Design

For the 31 localities, a database was obtained with the measurements of the 17 variables of the mesquite leaves, and the software of Statistical Package for Social Science (SPSS) version 15 for Windows was used to analyze the central tendency and dispersion, such as the average, standard deviation, values minimum and maximum and the coefficient of variation. A one-way analysis of variance and Tukey’s multiple comparison test of means [26] were performed to determine the difference and similarity between the localities, respectively, for the different parameters analyzed. A multivariate conglomerate analysis or cluster analysis was also used to integrate groups of sites that were associated with a taxonomic category assigned to mesquite and previously identified through traditional taxonomy. In relation to their groupings and their separation distances, dendrograms were generated that helped the formation of homogeneous and heterogeneous groups and the preliminary separation of the taxa, using a proximity matrix of Euclidean distances, where the number of clusters in each analysis was determined. The squared Euclidean distance between the groups (sites) was calculated using Formula (1), as follows:

$$\mathrm{Dij}=\sqrt{{\sum }_{\mathrm{k}=\mathrm{r}}^{\mathrm{n}}{\left(\mathrm{X}\mathrm{k}\mathrm{i}-\mathrm{X}\mathrm{k}\mathrm{j}\right)}^2}$$

(1)

where Dij is the distance between the groups (sites) i and j and Xki is the value of the variable Xk for group i.

The final discrimination for the separation of the taxa was completed by applying a discriminant analysis via Wilk’s lambda.

3. Results

3.1. Traditional Taxonomic Identification

After having carried out the traditional taxonomic identification with the keys of Johnston [8], five taxa were found (

Table 4. Taxa identified for the genus Prosopis in northeastern Mexico.

	NAME	ABBREVIATION
TAXA 1	Prosopis glandulosa var. glandulosa	P. g. var. g.
TAXA 2	Prosopis glandulosa var. glandulosa x Prosopis laevigata	P. g. var. g. x P. l.
TAXA 3	Prosopis laevigata	P. l.
TAXA 4	Prosopis laevigata x Prosopis glandulosa var. glandulosa	P. l. x P. g. var. g.
TAXA 5	Prosopis reptans var. cinerascens	P. r. var. c.

), of which there are two varieties (Prosopis glandulosa var. glandulosa and Prosopis reptans var. cinerascens), a pure or typical species (Prosopis laevigata) and two hybrids (Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa).

The pinnae of P. glandulosa var. glandulosa have curved leaflets of greater length and width, with greater interleaf spacing, but a lower number of pairs of leaflets per pinna. For its part, the hybrid P. glandulosa var. glandulosa x P. laevigata has straighter and smaller leaflets, but with a greater number of pairs of leaflets per pinna (Figure 3). In both taxa, the rachis reaches the same size. Morphologically, it is observed that P. glandulosa var. glandulosa x P. laevigata is an intermediate form between P. glandulosa var. glandulosa and P. laevigata, but more closely related to P. glandulosa, due to the size of the leaflets.

As for Prosopis laevigata, its pinna has leaflets with smaller interleaf spacing, smaller length and width in the leaflets and a very small rachis compared to the hybrid P. laevigata x P. glandulosa var. glandulosa (Figure 4). In both taxa, the same number of leaflets per pinna occurs. Morphologically, the taxon P. laevigata x P. glandulosa var. glandulosa is an intermediate or hybrid form between P. laevigata and P. glandulosa var. glandulosa, but more closely related to the taxon P. laevigata, due to the number of pairs of leaflets per pinna and its leaflet size.

Morphologically, the leaflets of Prosopis glandulosa are curved in the middle part of the leaflet, they are thin at the base and end at a point and they are larger in size compared to the other taxa. The leaflets of Prosopis laevigata are straight, wide from the base and do not end at a point and they are small in size compared to the other taxa. The leaflets of hybrids are very variable with respect to their size, but always straight in shape and with ranges between their two parental taxa. The taxon Prosopis reptans var. cinerascens is morphologically very different from the other taxa, since its leaflets are extremely small (Figure 5).

3.2. Analysis of Morphometric Parameters

For the taxa identified through traditional taxonomic keys, an analysis of morphometric parameters was carried out, as can be seen in

Table 5. Measurements of central tendency and dispersion of morphometric parameters in the 5 taxa of the 31 sampled localities.

Site	Locality—Municipality	Taxa *	Length		Length/Width Ratio		Spacing
			Middle	Upper	Middle	Upper
6	Ejido Puente del Río Salado—Anáhuac	P.g. var. g.	32.40 ± 0.43	26.87 ± 0.49	11.38 ± 0.11	9.69 ± 0.11	9.38 ± 0.19
23	Ejido las Presas—Lampazos	P.g. var. g.	31.50 ± 0.50	27.39 ± 0.50	11.15 ± 0.15	9.39 ± 0.13	8.58 ± 0.22
24	El Nogal—Anáhuac	P.g. var. g.	31.49 ± 0.32	28.18 ± 0.42	10.98 ± 0.15	9.76 ± 0.14	8.47 ± 0.14
25	Regantes 26—Anáhuac	P.g. var. g.	36.47 ± 0.36	30.77 ± 0.38	12.17 ± 0.13	10.58 ± 0.11	9.55 ± 0.17
1	Plan del Orégano—Melchor Ocampo	P.g. var. g X P.l.	26.02 ± 0.26	19.41 ± 0.31	9.88 ± 0.11	7.84 ± 0.10	7.47 ± 0.14
2	Colorados de Arriba—Vallecillo	P.g. var. g. X P.l.	29.15 ± 0.39	23.27 ± 0.38	11.08 ± 0.12	8.82 ± 0.11	7.91 ± 0.17
3	El Llano—Los Ramones	P.g. var. g X P.l.	24.69 ± 0.33	18.80 ± 0.35	10.23 ± 0.09	8.02 ± 0.09	6.77 ± 0.10
4	La Barretosa—Los Herrera	P.g. var. g. X P.l.	26.32 ± 0.36	19.90 ± 0.35	10.92 ± 0.12	8.10 ± 0.10	7.38 ± 0.12
5	Ejido Emiliano Zapata—Parás	P.g. var. g. X P.l.	30.15 ± 0.46	24.73 ± 0.39	10.78 ± 0.13	8.83 ± 0.10	8.38 ± 0.22
7	Loma Larga—Gral. Treviño	P.g. var. g. X P.l.	24.83 ± 0.28	19.90 ± 0.29	9.54 ± 0.09	7.78 ± 0.10	6.80 ± 0.10
9	Ejido El Álamo—Vallecillo	P.g. var. g. X P.l.	28.43 ± 0.41	21.16 ± 0.38	10.16 ± 0.11	8.04 ± 0.11	7.97 ± 0.15
10	Dulces Nombres—Pesquería	P.g. var. g. X P.l.	22.29 ± 0.29	17.26 ± 0.39	7.66 ± 0.09	6.04 ± 0.10	6.96 ± 0.22
11	Hacienda San Pedro—Gral. Zuazua	P.g. var. g. X P.l.	30.67 ± 0.38	22.40 ± 0.39	9.84 ± 0.12	7.79 ± 0.12	8.54 ± 0.16
12	Higueras—Higueras	P.g. var. g. X P.l.	28.27 ± 0.30	18.81 ± 0.36	10.27 ± 0.15	7.74 ± 0.14	7.59 ± 0.10
13	Los Pajaritos—Dr. González	P.g. var. g. X P.l.	25.64 ± 0.37	19.90 ± 0.39	9.81 ± 0.10	8.07 ± 0.10	7.34 ± 0.19
15	Loma la Parada—Marín	P.g. var. g. X P.l.	23.28 ± 0.31	18.62 ± 0.44	9.74 ± 0.09	7.90 ± 0.10	6.90 ± 0.14
16	El Bajío—Marín	P.g. var. g. X P.l.	23.86 ± 0.41	19.35 ± 0.39	8.69 ± 0.10	7.19 ± 0.10	7.27 ± 0.17
17	El Resumidero—Salinas Victoria	P.g. var. g. X P.l.	28.48 ± 0.30	22.79 ± 0.36	11.00 ± 0.14	8.92 ± 0.12	8.15 ± 0.17
18	El Puente—Salinas Victoria	P.g. var. g. X P.l.	27.36 ± 0.32	21.81 ± 0.34	10.04 ± 0.12	7.91 ± 0.09	7.94 ± 0.14
19	Rancho Gomas—Salinas Victoria	P.g. var. g. X P.l.	28.09 ± 0.43	23.90 ± 0.47	10.53 ± 0.09	8.89 ± 0.08	9.20 ± 0.24
26	Rancho la Ceja—Los Aldama	P.g. var. g. X P.l.	23.80 ± 0.31	19.43 ± 0.35	9.77 ± 0.13	7.98 ± 0.10	6.95 ± 0.15
21	Rancho La Bonanza—Gral. Terán	P. l.	9.19 ± 0.09	7.04 ± 0.09	5.31 ± 0.05	4.04 ± 0.05	2.24 ± 0.06
22	Rancho Nuevo—Gral. Terán	P. l.	11.45 ± 0.12	8.83 ± 0.11	5.87 ± 0.05	4.53 ± 0.04	2.73 ± 0.06
28	San Pedro de los Escobedos—Linares	P. l.	9.20 ± 0.13	7.07 ± 0.11	5.45 ± 0.05	3.99 ± 0.03	2.20 ± 0.05
30	Ejido las Puentes—Aramberri	P. l.	11.47 ± 0.19	9.21 ± 0.21	5.74 ± 0.07	4.46 ± 0.06	3.33 ± 0.08
8	Los Ébanos—Los Ramones	P.l. X P. g. var. g.	17.96 ± 0.29	13.41 ± 0.26	7.89 ± 0.07	5.76 ± 0.06	5.06 ± 0.12
14	Rancho el Recuerdo—Gral. Terán	P.l. X P. g. var. g.	16.69 ± 0.24	12.44 ± 0.25	7.60 ± 0.08	5.83 ± 0.08	4.98 ± 0.09
20	Km80 **—Los Ramones	P.l. X P. g. var. g.	19.36 ± 0.23	12.06 ± 0.21	7.08 ± 0.10	5.49 ± 0.09	4.72 ± 0.10
27	China—China	P.l. X P. g. var. g.	20.31 ± 0.23	15.07 ± 0.22	8.48 ± 0.06	6.27 ± 0.05	4.70 ± 0.09
29	San Ignacio de Texas—Galeana	P.l. X P. g. var. g.	14.79 ± 0.18	15.94 ± 0.21	8.56 ± 0.10	6.65 ± 0.08	4.90 ± 0.08
31	Km80 **—Los Ramones	P.r. var. c.	4.63 ± 0.04	2.76 ± 0.03	3.46 ± 0.04	1.59 ± 0.01	1.00 ± 0.02

* The abbreviations correspond to the initials of the species name: P. g. var. g. = Prosopis glandulosa var. glandulosa; P. g. var. g. X P. l. = Prosopis glandulosa var. glandulosa X Prosopis laevigata; P.l. = Prosopis laevigata; P.l X P.g. var. g. = Prosopis laevigata x Prosopis glandulosa var. glandulosa; P.r. var. c. = Prosopis reptans var. cinerascens.

. The variable middle leaflet length presented the highest average value in the northern part of the region, in the municipalities of Anáhuac and Lampazos and for the taxon Prosopis glandulosa var. glandulosa, the ranges of which spanned from 31.50 to 36.47 mm; meanwhile, in the central–southern and southern part, the lowest values of 9.19–11.47 mm were recorded; although Prosopis reptans var. cinerascens also has lower values, they could not be compared with any of the taxa, since morphologically it is very different from these. On the other hand, the hybrids from the central part presented intermediate values between the taxa from the northern and southern parts; that is, for the taxon Prosopis glandulosa var. glandulosa x Prosopis laevigata, their ranges are from 22.29 to 30.15 mm, more similar to those of the northern part (Prosopis glandulosa var. glandulosa) than to those of the southern part (Prosopis laevigata), and this happens inversely with the Prosopis laevigata x Prosopis glandulosa var. glandulosa, which has ranges from 14.79 to 20.31 mm.

In the case of the variable length of upper leaflets, Prosopis glandulosa var. glandulosa presents the highest average values with ranges of 26.87–30.77 mm, as this taxon is from the northern part of the region, while, in the southern and central–south part, in the municipalities of General Terán, Linares and Aramberri, the lowest average values for Prosopis laevigata were in the range of 7.07–9.21 mm, and, for the intermediate taxa Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa, the values were 17.26–24.73 mm and 12.44–15.94 mm, respectively. Finally, a particular and different morphology was presented for Prosopis reptans var. cinerascens, which recorded an average value of 2.76 mm.

For the variable of middle leaflet spacing, the highest average value was presented by Prosopis glandulosa var. glandulosa in locality 6 “Ejido Puente Río Salado”, in the north of the region, in the municipality of Anáhuac, while the lowest value was recorded in the central–southern part, in the municipality of Linares in locality 28 “San Pedro de los Escobedos”, with a value of 2.20 mm.

In the central part, average values of 6.77–9.20 mm and 2.20–3.33 mm were recorded for Prosopis glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa, respectively. Finally, for Prosopis reptans var. cinerascens, in the salty areas of the region, had average values of 1 mm of spacing in the middle leaflets, which were very different values compared to the other taxa.

For the length/width ratio of middle leaflets, the taxon Prosopis glandulosa presented the highest average value of 12.17 times as long as it was wide, in locality 25 “Los Regantes 26” of the municipality of Anahuac, in the north of the region. In the south–central part, in General Terán and in the locality 21 Rancho la Bonanza, the lowest value was recorded for the taxon Prosopis laevigata, which was 5.31 times as long as it was wide. On the other hand, in the center of the region, Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa presented values that were 7.66–11.08 and 7.08–8.56 times as long as they were wide, respectively. Finally, for Prosopis reptans var. cinerascens, an average value of 3.46 mm, as long as it was wide, was recorded, a value very different from the other taxa.

The length/width ratio of upper leaflets presented the highest average value in the north of the region, in locality 30 “Ejido Puentes” for the taxon Prosopis glandulosa var. glandulosa, which was 10.58 times as long as it is wide. In the south–central part in General Terán, in locality 21 “Rancho la Bonanza”, the lowest average value of 3.99 times as long as the width was recorded. In the central part, intermediate values of 6.04–8.92 and 5.49–6.65 times as long as the width were recorded in the taxa Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa. Finally, Prosopis reptans var. cinerascens presented a value of 1.59 times as long as it was wide, very different from the other taxa.

3.3. Analysis of Variance on the Variability of Morphometric Parameters

For all morphometric variables in the study, the analysis of variance (ANOVA) (

Table 6. ANOVA of the 17 study variables for the 31 sampled localities.

Variable	Source of Variation	Degrees of Freedom	Mean Squares	Values of F
Petiole length	Between groups	30	36,051.91	136.47 ***
	Within the group	6245	264.17
	Total	6275
Pinna width	Between groups	30	53,651.75	591.38 ***
	Within the group	6245	90.72
	Total	6275
Pinna number	Between groups	30	53,651.75	45.50 ***
	Within the group	6245	90.72
	Total	6275
Number of leaflets per pinna	Between groups	30	9203.88	373.45 ***
	Within the group	6245	24.64
	Total	6275
Rachis length	Between groups	30	89,633.42	248.71 ***
	Within the group	6245	360.39
	Total	6275
Lower leaflet length	Between groups	30	8445.18	374.70 ***
	Within the group	6245	22.54
	Total	6275
Lower leaflet width	Between groups	30	27.84	150.42 ***
	Within the group	6244	0.18
	Total	6274
Lower leaflet length/widthratio	Between groups	30	866.64	336.33 ***
	Within the group	6244	2.57
	Total	6274
Middle leaflet length	Between groups	30	13,108.13	655.73 ***
	Within the group	6245	19.99
	Total	6275
Middle leaflet width	Between groups	30	37.05	212.04 ***
	Within the group	6245	0.17
	Total	6275
Middle leaflets Length/width ratio	Between groups	30	956.32	440.65 ***
	Within the group	6245	2.17
	Total	6275
Upper leaflets length	Between groups	30	9484.38	425.94 ***
	Within the group	6245	22.26
	Total	6275
Upper leaflets width	Between groups	30	26.19	92.01 ***
	Within the group	6245	0.28
	Total	6275
Upper leaflet length/widthratio	Between groups	30	843.55	470.15 ***
	Within the group	6245	1.79
	Total	6275
Lower leaflet spacing	Between groups	30	494.72	259.41 ***
	Within the group	3107	1.90
	Total	3137
Middle leaflet spacing	Between groups	30	560.84	284.62 ***
	Within the group	3107	1.97
	Total	3137
Upper leaflet spacing	Between groups	30	693.32	251.14 ***
	Within the group	3107	2.76
	Total	3137

*** Highly significant difference (p = 0.000).

) shows highly significant differences (p = 0.000) between the localities, with F values ranging between 45.5 and 655.7, evidencing great variability in the morphometric parameters, both between the sites and within the same locality (between the trees). The high values of significant differences were recorded for the variables of middle leaflet length (F = 655.73), pinna width (F = 591.37), upper leaflet length/width ratio (F = 470.15) and upper leaflet length, while the low values of differences were obtained for the number of pinna (F = 45.5), width of upper leaflets (F = 92.0), length of petiole (F = 136.15) and width of lower leaflets (F = 150.42).

3.4. Tukey Multiple Comparison of Means

In the multiple comparison of means (Tukey) for the variable of middle leaflet length, 18 subgroups were formed for the 31 sampled locations with a 95% confidence interval. Localities 31, 21, 28, 22, 30, 29, 14, 27, 10, 13 and 25 formed an individual subgroup, while the remaining 20 localities belonged to two or more subgroups (

Table 7. Multiple comparison of Tukey means for the length of middle leaflets and upper leaflets in the 31 localities sampled in the semi-desert area of northeastern Mexico.

	Length of Middle Leaflets				Length of Upper Leaflets
Site	Locality	Taxa	Mean	Site	Locality	Taxa	Mean
31	Km 80 **	P.r. var. c.	4.63 A *	31	Km 80 **	P.r. var. c.	2.76 A *
21	Rancho La Bonanza	P. l.	9.18 B	21	Rancho La Bonanza	P. l.	7.03 B
28	San pedro de los Escobedos	P. l.	9.20 B	28	San pedro de los Escobedos	P. l.	7.06 BC
22	Rancho Nuevo	P. l.	11.45 C	22	Rancho Nuevo	P. l.	8.83 CD
30	Ejido las Puentes	P. l.	11.47 C	30	Ejido las Puentes	P. l.	9.20 D
29	San Ignacio de Texas	P.l. X P. g. var. g.	14.78 D	29	San Ignacio de Texas	P.l. X P. g. var. g.	12.06 E
14	Rancho el Recuerdo	P.l. X P. g. var. g.	16.68 E	14	Rancho el Recuerdo	P.l. X P. g. var. g.	12.44 E
8	Los Ébanos	P.l. X P. g. var. g.	17.96 EF	8	Los Ébanos	P.l. X P. g. var. g.	13.40 EF
20	Km 80 **	P.l. X P. g. var. g.	19.36 FG	20	Km 80 **	P.l. X P. g. var. g.	15.06 FG
27	China	P.l. X P. g. var. g.	20.30 G	27	China	P.l. X P. g. var. g.	15.93 GH
10	Dulces Nombres	P.g. var. g. X P.l.	22.29 H	10	Dulces Nombres	P.g. var. g. X P.l.	17.26 HI
15	Loma la Parada	P.g. var. g. X P.l.	23.28 HI	15	Loma la Parada	P.g. var. g. X P.l.	18.62 IJ
26	Rancho la Ceja	P.g. var. g. X P.l.	23.80 HI	3	El Llano	P.g. var. g. X P.l.	18.79 IJ
16	El Bajío	P.g. var. g. X P.l.	23.86 HI	12	Higueras	P.g. var. g. X P.l.	18.80 IJ
3	El Llano	P.g. var. g. X P.l.	24.69 IJ	16	El Bajío	P.g. var. g. X P.l.	19.35 J
7	Loma Larga	P.g. var. g. X P.l.	24.82 IJ	1	Plan del Orégano	P.g. var. g. X P.l.	19.41 JK
13	Los Pajaritos	P.g. var. g. X P.l.	25.64 J	26	Rancho la Ceja	P.g. var. g. X P.l.	19.43 JK
1	Plan del Orégano	P.g. var. g X P.l.	26.02 JK	7	Loma Larga	P.g. var. g. X P.l.	19.89 JK
4	La Barretosa	P.g. var. g. X P.l.	26.31 JK	13	Los Pajaritos	P.g. var. g. X P.l.	19.89 JK
18	El Puente	P.g. var. g. X P.l.	27.36 KL	4	La Barretosa	P.g. var. g. X P.l.	19.90 JK
19	Rancho Gomas	P.g. var. g. X P.l.	28.09 LM	9	Ejido El Álamo	P.g. var. g. X P.l.	21.15 KL
12	Higueras	P.g. var. g. X P.l.	28.27 LM	18	El Puente	P.g. var. g. X P.l.	21.80 LM
9	Ejido El Álamo	P.g. var. g. X P.l.	28.43 LM	11	Hacienda San Pedro	P.g. var. g. X P.l.	22.39 LMN
17	El Resumidero	P.g. var. g. X P.l.	28.48 LMN	17	El Resumidero	P.g. var. g. X P.l.	22.78 LMN
2	Colorados de Arriba	P.g. var. g. X P.l.	29.15 MNO	2	Colorados de Arriba	P.g. var. g. X P.l.	23.27 MNO
5	Ej. Emiliano Zapata	P.g. var. g. X P.l.	30.14 NOP	19	Rancho Gomas	P.g. var. g. X P.l.	23.89 NO
11	Hacienda San Pedro	P.g. var. g. X P.l.	30.66 OP	5	Ej. Emiliano Zapata	P.g. var. g. X P.l.	24.72 O
24	El Nogal—Anáhuac	P.g. var. g.	31.49 PQ	6	Ej. Puente del Río Salado	P.g. var. g.	26.86 P
23	Ejido las Presas—Lampazos	P.g. var. g.	31.49 PQ	23	Ejido las Presas	P.g. var. g.	27.38 P
6	Ej. Puente del Río Salado	P.g. var. g.	32.40 PQ	24	El Nogal	P.g. var. g.	28.17 P
25	Regantes 26—Anáhuac	P.g. var. g.	36.47 R	25	Regantes 26	P.g. var. g.	30.70 Q

Different letters indicate significant differences (p < 0.05). The abbreviations in the taxa correspond to the initials of the species name: P. g. var. g. = Prosopis glandulosa var. glandulosa; P. g. var. g. x P. l. = Prosopis glandulosa var. glandulosa x Prosopis laevigata; P.l. = Prosopis laevigata; P.l x P.g. var. g. = Prosopis laevigata x Prosopis glandulosa var. glandulosa; P.r. var. c. = Prosopis reptans var. cinerascens. * Significant difference (p = 0.000). ** Sites 20 and 31 corresponding to the same locality name, but with different species of Prosopis spp.

).

For the length of upper leaflets, there were seventeen subgroups, with localities 31, 21, 30, 29, 14, 16, 5, 6, 23, 24 and 25 forming an independent subgroup, while the remaining 20 localities belonged to two or more subgroups (Table 7).

In middle leaflet spacing, eleven subgroups were formed, where localities 31, 28, 21, 30, 20, 29, 27, 14, 8, 3, 7, 24, 6 and 25 constituted an individual subgroup, and the remaining 17 localities belonged to two or more subgroups (

Table 8. Multiple comparison of Tukey means for the variables of spacing of middle leaflets, length/width ratio of middle leaflets and length/width ratio of upper leaflets in the 31 sampled locations.

Spacing of Middle Leaflets			Length/Width Ratio of Middle Leaflets			Length/Width Ratio of Upper Leaflets
Site	Locality	Mean	Site	Locality	Mean	Site	Locality	Mean
31	Km80 **	1.00 A *	31	Km80 **	3.45 A *	31	Los Ramones	1.58 A *
28	San pedro de los Escobedos	2.19 B	21	Rancho La bonanza	5.30 B	28	Linares	3.98 B
21	Rancho La bonanza	2.24 B	28	San pedro de los Escobedos	5.45 BC	21	General Terán	4.03 BC
22	Rancho Nuevo	2.72 BC	30	Ejido las Puentes	5.73 BC	30	Aramberri	4.46 BC
30	Ejido las Puentes	3.33 C	22	Rancho Nuevo	5.86 C	22	General Terán	4.52 C
20	Km 80 **	4.70 D	29	San Ignacio de Texas	7.07 D	29	Galeana	5.49 D
29	San Ignacio de Texas	4.71 D	14	Rancho el Recuerdo	7.60 DE	8	Los Ramones	5.75 DE
27	China	4.90 D	10	Dulces Nombres	7.65 E	14	General Terán	5.82 DEF
14	Rancho el Recuerdo	4.97 D	8	Los Ébanos	7.89 E	10	Pesquería	6.03 EF
8	Los Ébanos	5.05 E	20	Km80 **	8.48 F	20	Los Ramones	6.27 FG
3	El Llano	6.77 E	27	China	8.56 F	27	China	6.65 G
7	Loma Larga	6.79 E	16	El Bajío	8.69 F	16	Marín	7.18 H
15	Loma la Parada	6.89 EF	7	Loma Larga	9.53 G	12	Higueras	7.74 I
26	Rancho la Ceja	6.94 EF	15	Loma la Parada	9.73 GH	7	General Treviño	7.78 I
10	Dulces Nombres	6.96 EF	26	Rancho la Ceja	9.77 GH	11	General Zuazua	7.78 I
16	El Bajío	7.77 EFG	13	Los Pajaritos	9.80 GH	1	Melchor Ocampo	7.84 I
13	Los Pajaritos	7.33 EFG	11	Hacienda San Pedro	9.83 GH	15	Marín	7.89 I
4	La Barretosa	7.38 EFG	1	Plan del Orégano	9.88 GH	18	Salinas Victoria	7.90 I
1	Plan del Orégano	7.46 EFGH	18	El Puente	10.03 GHI	26	Los Aldama	7.98 I
12	Higueras	7.59 FGH	9	Ejido El Álamo	10.15 HI	3	Los Ramones	8.02 I
2	Colorados de Arriba	7.90 GHI	3	El Llano	10.22 HIJ	9	Vallecillo	8.04 I
18	El Puente	7.94 GHI	12	Higueras	10.26 HIJ	13	Doctor González	8.07 I
9	Ejido El Álamo	7.97 GHI	19	Rancho Gomas	10.52 IJK	4	Los Herreras	8.10 J
17	El Resumidero	8.14 GHI	5	Ej. Emiliano Zapata	10.77 JKL	2	Vallecillo	8.81 J
5	Ej. Emiliano Zapata	8.37 HI	4	La Barretosa	10.91 KLM	5	Parás	8.83 J
24	El Nogal	8.47 I	24	El Nogal	10.97 KLM	19	Salinas Victoria	8.88 JK
11	Hacienda San Pedro	8.53 IJ	17	El Resumidero	11.00 KLM	17	Salinas Victoria	8.91 JK
23	Ejido las Presas	8.58 IJ	2	Colorados de Arriba	11.07 KLM	23	Lampazos	9.38 KL
19	Rancho Gomas	9.20 JK	23	Ejido las Presas	11.14 LM	6	Anáhuac	9.69 L
6	Ej. Puente del Río Salado	9.37 K	6	Ej. Puente del Río Salado	11.37 M	24	Lampazos	9.76 L
25	Anáhuac	9.54 K	25	Regantes 26-Anáhuac	12.16 N	25	Regantes 26 - Anáhuac	10.57 M

Different letters indicate significant differences (p < 0.05). * Significant difference (p = 0.000). ** Sites 20 and 31 corresponding to the same locality name, but with different species of Prosopis spp.

).

Regarding the length/width ratio of middle leaflets, fourteen subgroups were formed, of which the localities 31, 21, 22, 29, 10, 8, 20, 27, 16, 7, 6 and 25 are an individual subgroup, while the remaining 20 localities belong to two or more subgroups (Table 8).

The length/width ratio of upper leaflets, for its part, recorded thirteen subgroups, where the localities 21, 30, 8, 14, 10, 20, 19, 17 and 23 belonged to two or more subgroups, while the remaining 22 localities formed individual subgroups (Table 8).

3.5. Grouping of Taxa Using Cluster Analysis

In the cluster analysis, the 17 study variables were used for the 31 localities, and between these variables, an exhaustive combination was made to form phenograms and select those that best group the taxa.

Figure 6 shows the homogeneity of the data, considering the red line of Euclidean distance <1, while the blue line of Euclidean distance <5 shows the formation of the following six groups: A, B, C, D, E and F.

Group A is made up of the “pure taxon” (P. glandulosa var. glandulosa), with sites 6, 23 and 24, and the “intermediate or hybrid” taxon (P. glandulosa var. glandulosa x P. laevigata), with the sites 2, 5, 9, 11, 12, 17, 18 and 19. Group B is formed by the intermediate taxon (P. glandulosa var. glandulosa x P. laevigata) with sites 1, 3, 4, 7, 10, 13, 15 and 26. These two groups, A and B, have a degree of kinship, except for sites 6, 23, 24 and 10. In addition, these two groups have a greater number of hybrid sites or localities than pure ones. Group C is made up only of site 25, and although it is a separate group, it is related to the glandulosa (groups A and B). For group D, sites 8, 14, 20, 27 and 29 enter with the intermediate taxon P. laevigata x P. glandulosa var. glandulosa, and, for group E, there is the taxon P. laevigata, with sites 21, 22, 28 and 30. For its part, group F is made up of the taxon P. reptans var. cinerascens, and is very different morphologically from the others, but with measurements close to the taxon P. laevigata (groups D and E).

In Figure 7, drawing a red line with an Euclidean distance of <1, the homogeneity of the data is shown, while, in the blue line with Euclidean distance of <5, the following six groups are formed: A, B, C, D, E and F.

Groups A and B contain mostly the taxon P. glandulosa var. glandulosa x P. laevigata at sites 1, 2, 3, 4, 5, 7, 9, 11, 12, 13, 15, 16, 17, 18, 19 and 26, except for sites 20, 27 and 10, which are from the taxon P. laevigata x P. glandulosa var. glandulosa; both groups are intermediate forms or hybrids.

Group C is clearly defined by the taxon P. glandulosa var. glandulosa with sites 6, 23, 24 and 25; for group D, there is the intermediate taxon P. laevigata x P. glandulosa var. glandulosa with sites 8, 14 and 29; for group E, there is the taxon P. laevigata with sites 21, 22, 28 and 30. These, as well as group C, are clearly defined because they are the pure or typical species. Finally, group F is made up only of the taxon P. reptans var. cinerascens at site 31.

In Figure 8, the homogeneity of the data can be observed, according to the red line of Euclidean distance <1, while the blue line of Euclidean distance <5 shows the following five groups: A, B, C, D and E.

Group A is formed by the taxon P. laevigata x P. glandulosa var. glandulosa, with sites 8, 14, 20, 27 and 29, and group B is formed by the pure or typical taxon P. laevigata, with sites 21, 22, 28 and 30. Both groups are clearly defined, as is group C, which is formed only by the taxon P. reptans var. cinerascens.

Groups D and E mostly have the hybrids P. glandulosa var. glandulosa x P. laevigata, except for sites 10, 11, 23 and 24 which overlap with the previous group.

In Figure 9, the homogeneity of the data is shown by the red line with an Euclidean distance of <1, while the blue line with an Euclidean distance of <5 shows the formation of the following four groups: A, B, C, and D.

In group A, the taxon P. glandulosa var. glandulosa x P. laevigata was at 15 sites, except for sites 6, 23, 24 and 25, which belonged to the taxon P. glandulosa var. glandulosa. Group B presents the taxon P. laevigata x P. glandulosa var. glandulosa except for site 16, which belongs to the taxon P. glandulosa var. glandulosa x P. laevigata.

Group C is well defined by the taxon P. laevigata and does not overlap with any taxa as in the previous groups (A and B). In group D, there is only the taxon P. reptans var. cinerascens, which is also clearly defined from the other groups.

In Figure 10, the red line with a Euclidean distance of <1 shows the homogeneity of the data, and the blue line with a Euclidean distance of <5 shows the formation of the following five groups: A, B, C, D and E.

Group A is made up of the taxon P. glandulosa var. glandulosa x P. laevigata with 12 localities, except locality 27, which overlapped with this group, and belongs to the taxon P. laevigata x P. glandulosa var. glandulosa.

In group B are the taxa P. glandulosa var. glandulosa x P. laevigata and P. glandulosa var. glandulosa, with sites 2, 5, 17 and 19 and 6, 23, 24 and 25, respectively. Group C is clearly defined by the taxon P. laevigata. Group D is defined by sites 10, 20, 8, 14 and 29; and, finally, group E is made up only of the taxon P. reptans var. cinerascens.

3.6. Discriminant Analysis

For the discriminant analysis, the database of 31 localities was used and each locality was labeled with its respective taxa, according to the traditional taxonomic identification carried out in the present study.

In

Table 9. Wilk’s lambda Values.

Contrast of Functions	Wilk’s Lambda	Chi-Squared	Degrees of Freedom	Sig.
1 to 4	0.196	5105.635	20	0.000
2 to 4	0.895	345.854	12	0.000
3 to 4	0.975	79.802	6	0.000
4	0.997	10.851	2	0.000

, showing Wilks’ lambda for the contrast of the functions, it appears that taxa 1 to 4 have lambda values (0.196) closer to zero than one, and therefore, the groups are different; also, the chi-squared values (4899.423), plus the degrees of freedom (20), give a value of zero, so there is a highly significant difference (p = 0.000) between the groups or between the taxa. For the other contrasts of functions such as between taxa 2 to 4, taxa 3 to 4 and in taxon 4, the lambda values are close to one, indicating that they are equal to each other, but the chi-squared values plus the degrees of freedom give a significant difference between taxa or between groups.

The predicted membership groups are presented in

Table 10. Membership or kinship groups among the 5 taxa.

Taxa		Predicted Membership Group *					Total
		1	2	3	4	5
Count	1	194	205	0	1	0	400
	2	110	1371	0	123	0	1604
	3	0	2	422	30	0	454
	4	0	145	45	390	0	580
	5	0	0	3	0	97	100
%	1	48.5	51.3	0.0	0.3	0.0	100
	2	6.9	85.5	0.0	7.7	0.0	100
	3	0.0	0.4	93.0	6.6	0.0	100
	4	0.0	25.0	7.8	67.2	0.0	100
	5	0.0	0.0	3.0	0.0	97.0	100

* 78% of original grouped cases were correctly classified.

, which indicates that 78.8% of the grouped cases have been correctly classified for the five taxa (P. laevigata, P. laevigata x P. glandulosa var. glandulosa, P. glandulosa var. glandulosa x P. laevigata and P. reptans var. cinerascens). For taxon 1 (P. glandulosa var. glandulosa), 48.5% of the cases belong to the same taxon, and 51% belong to taxon 2 (P. glandulosa var. glandulosa x P. laevigata), so the degree of grouping of the cases is very similar. Regarding taxon 2 (P. glandulosa var. glandulosa x P. laevigata), a percentage of 85.5% belong to the same taxon; 6.9 belong to taxon 1 and 7.7% belong to taxon 4, although this taxon presents a high % value of the correctly classified cases, which is due to the greater number of localities or cases it had with respect to the other taxa, confirming that, in most cases, they are intermediate forms. For its part, taxon 3 (P. laevigata) has a very high percentage of 93% and is clearly defined among the same group. For taxon 4 (P. laevigata x P. glandulosa var. glandulosa), 67.2% of the cases belong to the same group or taxon, but there is a predicted 18.3% that belong to taxon 2 (P. glandulosa var. glandulosa x P. laevigata). Finally, there is taxon 5 (P. reptans var. cinerascens), which corresponds to the highest value with 97% and which is clearly defined and differentiated from the other groups.

The scatter diagram (Figure 11) is constructed based on the canonical discriminant functions not typified in the group means, and the cases are identified by taxa. Taxon 1 (P. glandulosa var. glandulosa) and taxon 2 (P. glandulosa x P. laevigata) overlap, because the values of the centroids or the values of the means of these taxa in the discriminant functions are so close that they are not completely separated or that they share morphological characteristics with each other. The same happens with taxon 4 (P. laevigata x P. glandulosa var. glandulosa), which overlaps with taxon 2 (P. glandulosa var. glandulosa x P. laevigata) and taxon 3 (P. laevigata), although the latter is very distant from taxa 1, 2 and 5. Taxon 5 (P. reptans var. cinerascens) is clearly defined from the other taxa, because its average is very distant from the other taxa and because it is also very different morphologically.

The present scatter diagram seems somewhat fruitless, since there is overlap between groups or taxa from 1 to 4, except for 5; this overlap does not discriminate or separate the groups, as would have been expected, but, nevertheless, it shows that there are intermediate forms (hybridization). Although the distances between the mean values of the centroids are very close, there is a very defined nucleus, which prevents the consideration of the taxa as equal.

Figure 12 shows the location of the 31 localities with their corresponding taxon, according to the interpretation of the cluster analysis and the discriminant analysis.

4. Discussion

The variables length of middle leaflets and length/width ratio of middle leaflets present average values of 31.49 to 36.47 mm and 10.98 to 12.17 mm, respectively, for the variety Prosopis glandulosa var. glandulosa, which fall within the morphometric ranges of the leaves as reported by Ábrego-Rodríguez [27], who mentions that Prosopis glandulosa has the following two varieties: P. glandulosa var. glandulosa, with leaves measuring 30–35 mm long, eight to fifteen times as long as they are wide, and the variety P. glandulosa var. torreyana, which is characterized by leaves measuring 25–30 mm long, five to eight times as long as they are wide. On the other hand, the intermediate or hybrid taxon of the present study Prosopis glandulosa var. glandulosa x Prosopis laevigata recorded average values (22.29–30.67 mm and 7.66–11.08 mm), similar to the taxon P. glandulosa var. torreyana described by the same author in the previously mentioned variables (length of middle leaflets and length/width ratio of middle leaflets); the length does agree, but the length/width relationship does not, so the values obtained in the present study are values that correspond to a hybrid, since they are intermediate measures.

In a study comparing the measurements of foliar characteristics of some potential hybrids whose origin could have been generated by P. glandulosa var. torreyana and P. laevigata, Galindo-Almanza [28] found that, when crossing P. glandulosa var. torreyana and P. laevigata, for the variables length of the middle leaflets and interleaf spacing of middle leaflets, the hybrid had a leaflet length of 15 to 17 mm and an interleaf distance of 6 to 6.5 mm, similar to the values of 14.79–20.31 mm and 4.70–5.06 mm recorded for the hybrid P. laevigata x P. glandulosa var. glandulosa of the present study.

On the other hand, the same author reported values of 3 to 15 mm for leaflet length and 2.5–5 mm for interleaf distance for P. laevigata, confirming the average values of the present study for P. laevigata, which falls within the ranges reported for length of middle leaflets, length of upper leaflets and spacing of middle leaflets.

There are other hybrids in mesquite species, as mentioned by Burkart [6], Graham [7], Johnston [8] and Leakey and Last [29], based only on morphological characteristics, or, as indicated by Jiménez [30] and López [31], based on the enormous variability of some foliar morphometric parameters and on the taxonomic identification of mesquite in the center of the state of Nuevo León, as this hybrid is an intermediate form between the species Prosopis glandulosa and Prosopis laevigata. In addition, they mention the existence of Prosopis glandulosa var. torreyana in low densities, which could also be that hybrid from the center of the state. With respect to hybridization, the results obtained show that, in the variables of length of middle leaflets, length of upper leaflets, spacing of middle leaflets, length/width ratio of middle leaflets and length/width ratio of upper leaflets, there is no variety named torreyana, but there is an intermediate form between the taxa P. glandulosa var. glandulosa and P. laevigata, although a variety can arise spontaneously in nature and a subspecies is a very similar concept to a variety. Morphologically, it is not defined in any of the variables mentioned to consider it as such, so this enormous morphological variability and malformations in mesquite leaves and their geographical distribution in a contact zone between the two parental taxa (Prosopis glandulosa var. glandulosa and Prosopis laevigata) indicate that it is an hybrid and not a variety or subspecies, as it differs from the parent taxa, which are clearly defined in the selected variables from a morphological point of view and are geographically found in opposite parts of the state, namely in the northern and southern parts.

According to Rzedowski [32], the varieties Prosopis glandulosa var. glandulosa and Prosopis glandulosa var. torreyana are characteristic of Texas and are distributed in Nuevo León, Tamaulipas and eastern Coahuila, and it is common to recognize these two varieties. The present study, with the support of statistical analyses and taxonomic identification based on the keys of Johnston [8], confirms the existence of the variety Prosopis glandulosa var. glandulosa in a very defined way for the state of Nuevo León, which is located in the north and northwest of the state, in the municipalities of Lampazos and Anáhuac. However, this is not the case with respect to the variety Prosopis glandulosa var. torreyana, since the study carried out reveals that this variety is a hybrid, presenting intermediate characteristics between Prosopis glandulosa var. glandulosa and Prosopis laevigata, with the leaflet length falling in the range of the variety Prosopis glandulosa var. glandulosa and the number of leaflets per pinna corresponding to the second species, Prosopis laevigata.

On the other hand, the present study records the distribution of Prosopis laevigata in the southern part of the state of Nuevo León, in the municipalities of General Terán, Linares and Galeana and, in addition, there is a fairly wide strip of hybridization between the populations of Prosopis glandulosa var. glandulosa and Prosopis laevigata, in the center of the state of Nuevo León and covering the municipalities of Los Ramones, part of Gral. Terán, China, Aramberri, Melchor Ocampo, Vallecillo, Los Herrera, Parás, Gral. Treviño, Gral. Zuazua, Pesquería, Higueras, Doctor González, Marín, Los Aldama and Salinas Victoria. This coincides with Jiménez [31], who reported that the species Prosopis laevigata is distributed in southern Nuevo León, and that, in turn, its distribution area makes contact with Prosopis glandulosa var. glandulosa, creating a strip of trees with intermediate characteristics in the center of the state.

In his study of mesquites in North America, Johnston [8] reported data obtained from leaves of herbaria specimens, although they were not subject to statistical analysis. However, he made a species key, and the ranges he managed are very similar to the results of this work, and it is the most accepted classification in the national territory. On the other hand, the same author, as well as Galindo [33] and Solbring et al. [34], mentioned that flowering periods, sympatry, the non-existence of geographical barriers and the little discrimination of pollinating insects increase hybridization opportunities for this genus. In our study, by geographically observing the sampling sites, it can be established that there is close contact between the pure populations, and we can assume that the non-existence of important geographical barriers that prevent overlap between these populations enable the exchange of genes (introgression). In addition to the climatic conditions, the close flowering periods, the winds coming from the gulf and the non-discrimination of pollinating insects as mentioned by the authors, the above are the possible causes of hybridization of this genus in the state of Nuevo León. Although this study does not address the causes of hybridization mentioned, this hybridization is reflected in the results obtained, in the existence of a contact zone or a hybridization strip in the central part of the state, where the morphology of the leaves of the taxa of Prosopis spp. demonstrates intermediate characteristics between the parental taxa.

5. Conclusions

Based exclusively on the morphological characters evaluated in this work, it can be said that there are five taxa of mesquite in the 31 localities sampled in the Mexico’s semi-desert area, of which there are two varieties (Prosopis glandulosa var. glandulosa, Prosopis reptans var. cinerascens), two hybrids (Prosopis glandulosa var. glandulosa x Prosopis laevigata, Prosopis laevigata x Prosopis glandulosa var. glandulosa) and a pure or typical species (Prosopis laevigata). Both the traditional taxonomic classification and the cluster analysis, as well as the discriminant analysis, reflect the same taxa (Prosopis glandulosa var. glandulosa, Prosopis laevigata and Prosopis reptans var. cinerascens) and the existence of hybridization between Prosopis glandulosa var. glandulosa and Prosopis laevigata. However, the study of other characteristics, such as molecular or cytological markers, could yield different results and would possibly lead to the proposal of other hypotheses to explain the observed variation.

On the other hand, the length, width and interleaf spacing, as well as the length/width ratio of the lower, middle and upper leaflets, tend to decrease from north to south, but the number of leaflets per pinna increases in the same direction.

In the north of the region, in the municipalities of Anáhuac and Lampazos, we can find the variety P. glandulosa var. glandulosa. The central zone of the state of Nuevo León is dominated by the taxa Prosopis glandulosa var. glandulosa x Prosopis laevigata and Prosopis laevigata x Prosopis glandulosa var. glandulosa, and there is a strip or zone of hybridization, marked by the municipalities of Los Ramones, part of Gral. Terán, China, Aramberri, Melchor Ocampo, Vallecillo, Los Herrera, Parás, Gral. Treviño, Gral. Zuazua, Pesquería, Higueras, Doctor González, Marín, Los Aldama and Salinas Victoria. In the saline soils of the center of the state, the taxon Prosopis reptans var. cinerascens is found. To the south, the taxon Prosopis laevigata is found in the municipalities of General Terán, Linares and Galeana.