Materials and methods

Study area. The Baja California Tropical Deciduous Forest is located at the southern top of the Baja California peninsula, in the Cape Region, between 500–1,000 m; the annual average rainfalls and temperatures are 200–400 mm and 22–24 °C, respectively (^{González-Abraham et al. 2010}, ^{León de la Luz et al. 2012}). The flora of this area is represented by species such as Bursera epinnata, Colubrina viridis, Jatropha cinerea, Karwinskia humboldtiana, Lysiloma candidum, Lysiloma microphyllum, and Stenocereus thurberi (^{León de la Luz et al. 2012}). Our research was carried out in the municipality of La Paz and it included localities in the “Ejido Álvaro Obregón”, community of “San Blas-Los Divisaderos”, and “Sierra Cacachilas” (Figure 1). The studied populations in these communities are scattered in small settlements from 1 to 30 houses, but for our analysis were included only the smallest (under 10 houses) by considering they are more related with wildness in contrast with the formal villages. In selected communities, people call themselves ranchers (rancheros, in Spanish), defined as people living in rural communities (ranches) in which livestock farming is their main economic activity, complemented with cheese works, home gardens, construction, public services, and sale of non-timber forest products (e.g. fruits Stenocereus). The ranches appeared just after retirement of the Jesuit missions in 1768 which started with small scale agriculture and livestock farming; later, immigrant ranchers from northwest Mexico had facilities for land tenure and adapted their knowledge to the semi-desert environment of this region, rising the current culture of the Southern Baja California ranchers; there is no native indigenous people in this area because the native hunter-gatherer Prericúe and Guaycura disappeared after the conquest (^{Urciaga-García 2008}). The selected area also has one of the largest concentrations of scattered settlements in the Cape Region and in the Tropical Deciduous Forest of Baja California peninsula.

Figure 1 Localities of the Cape Region where the study was conducted: Ejido Alvaro Obregón (1), San Blas-Los Divisaderos (2), and Sierra Cacachilas (3). 

Field work. Semi-structured interviews were conducted from March to November 2015. A total of 53 people were interviewed: 27 men and 26 women. The mean age for men participants was 49 ± 16 (from 20 to 81) and 58 ± 14 (from 34 to 86) for women. A maximum of two participants per home were included, mostly by individual interviews. A base of three simple questions was used:

Previously, a checklist of edible plants was created with local names and pictures from the local literature (^{Rebman & Roberts 2012}, ^{León de la Luz et al. 2014}); this checklist was used when the interviewed did not mention any new species. After the interview, volunteer participants were involved in field work for in situ identification of plant species and to get additional information, photographs and/or samples for the herbarium. The voucher specimens were identified and deposited in the herbarium “Annetta Mary Carter” (HCIB) of the Centro de Investigaciones Biológicas del Noroeste (CIBNOR).

Food Significance Index. Only those species with three or more mentions were considered. A simplified Food Significance Index (FSI) applied in a previous study (Pío-León J.F., pers. com., submitted for publication) was used and defined by the following equation:

where CV and CrV are Cultural Value and Crop Value sub-indices, respectively.

Cultural Value (CV) is calculated as CV = ci + ca + pr. Where: ci is the citation index defined as the number of participants who cited a species divided by the total number of participants; ca is the food category of the specie and represents its role in diet; and pr is the preference rank of the plant. In relation with ca, four categories were considered: staple foods, complements, seasoning and teas, and minor foods. Staple foods provide most of the calories and vegetal proteins to diet (e.g. cereals, legumes, potatoes); complements are important foods, but they are not the main sources of protein and calories (fruits, vegetables, and almond-like seeds); seasoning and teas add desirable sensory features to foods or water, but do not represent significant amounts of calories or other nutrients (such as pepper, oregano, green tea); minor foods are small fruits consumed occasionally in field, usually as a candy, and hardly became regular eating or take away hunger. ca values are 1.0 for staple foods, 0.75 for complements, 0.25 for seasoning and teas, and 0.1 for minor foods. On the other hand for the preference parameter (pr), each participant listed three species of their choice and the parameter was calculated as pr = [(p1/t)/1] + [(p2/t)/2] + [(p3/t)/3], where, for a specific plant, p1, p2, and p3 are the number of people who cite it as their first, second, or third choice; t is the total number of participants. Participants could list more than one species in each position when they so desired, such as two plants in p1, one in p2 and two in p3. The maximum theoretical value of pr for a species is 1 (preferred food by all participants) and the minimum is 0 (not cited in any position).

The Crop Value sub-index (CrV) was determined as CrV = pl + pp + neo. pl is the postharvest shelf life of the food under ambient conditions; three levels of pl were defined: non-perishable foods (e.g. almond-like seeds, dry leaves) (1 point value), medium term perishable (e.g. potatoes, semi-dry fruits) (0.5 points value), and highly perishable (fresh fruits and vegetables). pp is processed product with market potential; it is considered when the food item of the species is transformed into a marketable or potential marketable product, such as fermented beverage, jam or prickle. Those fresh foods which already have a commercial value were included as pp. Foods having a pp attribute were assigned a value of 1, otherwise it is 0. nep is the number of edible organs/parts of the plant. A number of maximum four edible organs were established: 1) underground organs, 2) leaves of branches, 3) fruits, and 4) seeds. For every edible part a value of 0.25 was given.

Sub-indices were multiplied in order to increase the variability of FSI and to identify differences between species. In a previous work (Pío-León J.F., pers. com., submitted for publication), a Management Value was included as a third sub-index; however, people in the interviewed communities have done few efforts for management or domestication of the wild species. The use of the Management Value sub-index is recommended for other areas where the researcher detects the management/ domestication of wild plant species.

Salience Index. Salience Index (SI) was proposed by ^{Smith (1993)} to analyze free list surveys and was recently recommended by ^{Kujawska & Łuczaj (2015)} to select the most important wild edible species. Results of the SI were compared with those of the FSI. For each species, Salience Index is defined as the average of the following ratio in all the free list interviews:

Salience Index = (total number of species in the interview – rank order position of the species (starting from 0 for the first species)) / total number of species cited in the interview.

Selection of the priority species. The priority species for food and conservation purposes were designated according to a selection system that promotes the culinary diversity. First, species were grouped based on the type of food: 1) seeds, 2) fruits and candy-type foods, 3) seasoning or tea, and 4) vegetables. Species in each group were sorted in descending order according their FSI and the lowers limits of the first, second, and third quartile were calculated, corresponding to the lower limits of FSI for the species in the Priority Group 1, Priority Group 2, and Priority Group 3, respectively. Equations were formulated as follow:

FSI of Priority Group 1 = >FSI_L + [(FSI_H – FSI_L) × 0.75]

FSI of Priority Group 2 = > FSI_L + [(FSI_H – FSI_L) × 0.5] – FSI of Priority Group 1

FSI of Priority Group 3 = > FSI_L + [(FSI_H – FSI_L) × 0.25] – FSI of Priority Group 2

FSI_L and FSI_H are the lowest and the highest values of FSI for each group of food, respectively. Using this selection system the priority group 1 was ensured to include at least one species of each group of foods. We used this selection system to avoid the likely bias of the indices for some kind of food. The same procedure was applied for the values of SI instead of FSI.

Results

Fifty-one taxa of wild edible plants were recorded with more than three mentions, which represent about 8 % of the total taxa for the deciduous forest in the Baja California peninsula (^{León de la Luz et al. 2012}). Families with higher number of taxa were Cactaceae (11), Fabaceae (7), Solanaceae (3), and Verbenaceae (3) (Table 1). Out of the 51 taxa, 25 were used as fruits, nine as vegetables, four as almonds-like or nuts, four as recreational tea, eight as candy, and two as seasonings (Table 2). Only the acorns of Quercus brandegeei were identified as staple food; at its harvesting season (November-December), acorns are mostly consumed as atole, a kind of thick drink prepared by mixing and boiling the toasted and ground acorns with milk. People usually drink atole during breakfast or dinner. Atole is also commonly prepared with the seeds of Quercus tuberculata or Cnidoscolus maculatus; however, these are less preferred and mostly consumed as almond-like seeds.

The 10 species with the highest FSI included a high culinary diverse group of food: two seeds (Quercus brandegeei and Cnidoscolus maculatus), four fruits (Ficus petiolaris, Stenocereus thurberi, Stenocereus gummosus, and Cyrtocarpa edulis), one recreational tea (Turnera diffusa), two seasoning (Lippia palmeri and Capsicum annuum), and one vegetable (Pithecellobium dulce). Pithecellobium dulce arils are commonly consumed as fruits (raw fresh arils) in other parts of Mexico and Asia (^{Parrotta 1991}, ^{Monroy & Colín 2004}), but people from this region used them as vegetables accompanying cooking. For this purpose, fresh arils were blanched and mixed with meat or other vegetables. Differences were detected among SI and FSI; in general, the highest values were for seeds (FSI) and fruits (SI); the top 10 species with the highest SI included a less culinary diverse group than FSI (Table 2). Salience Index was more related with the preference parameter (pr) and the Cultural Value (CV) (most popular species) than with FSI or the Crop Value (CrV).

As processed products with marketing potential, ranchers produced: marmalade/jams from eight fruits (Ficus petiolaris, Stenocereus thurberi, Stenocereus gummosus, Cyrtocarpa edulis, Diospyros californica, Pachycereus pecten-aboriginum, Pachycereus pringlei and Pithecellobium dulce) and the stems of Ferocactus townsendianus; wine or liqueur from the fruits of S. thurberi, S. gummosus, C. edulis, and Vitis peninsularis; prickles from C. edulis, Matelea cordifolia, and Capsicum annuum; coffee-like substitutes and powder to make atole from the seeds of Quercus brandegeei, Cnidoscolus maculatus, and Quercus tuberculata; and three are sold as dehydrated products for tea (Turnera diffusa) and seasoning (Lippia palmeri and C. annuum).

The fresh fruits of Stenocereus thurberi and S. gummosus (“pitayas”) are the best-selling products throughout the year. Stenocereus thurberi fruits has the highest economic value; these fruits are smaller, sweeter, and their harvesting season is shorter (July-August) than those of S. gummosus (August-December). There are two main S. thurberi phenotypes (red and white fruits), both of them are sold as fresh fruits, but marmalade is produced only from the red ones. Many people of the studied communities work in S. thurberi harvesting, but only some of them are involved in the commercialization in nearby urban centers, directly without going through supermarkets. By contrast, S. gummosus has only one main phenotype (red fruits), its postharvest shelf life is few days longer and they are expended in the supermarkets of the nearby urban centers.

Men cited more species than women. The age did not affect the average number of listed species by men, but the younger women cited less than the older ones (Table 3).

Selection of the priority species. Ten species were included in the Priority Group 1 using FSI values: one almond (Quercus brandegeei) (source of proteins and lipids), four fruits (Ficus petiolaris, Stenocereus thurberi, Stenocereus gummosus, and Cyrtocarpa edulis) (source of sugars, fiber, vitamins and minerals), one tea (Turnera diffusa) (source of secondary metabolites and minerals), two seasonings (Capsicum annuum and Lippia palmeri) (source of secondary metabolites), and two vegetables (Pithecellobium dulce and Matelea cordifolia) (source of fiber, vitamins, and minerals) (Table 4). The seeds of Cnidoscolus maculatus (Priority Group 2) had higher FSI value than most plant species in the Priority Group 1; however, they are the second choice after those of Quercus brandegeei. Foods in Priority Group 3 were diverse; they were the option by scarcity of foods of the Priority Group 1 or 2. For example, the consumption of Pachycereus pecten-aboriginum was more important in seasons when fruits of Stenocereus thurberi and S. gummosus were scarce. The acorns of Quercus tuberculata are very known by the ranchers, but this species is distributed only in the highest mountains and is consumed by few people. Another example is Lysiloma microphyllum, its gum is highly consumed by ranchers, but they prefer the tastier and more nutritious fruits of the Priority Group 1.

The use of SI values generated a similar pattern of diversity in grouping. The Priority Group 1 included almost the same species, but Cyrtocarpa edulis, Lippia palmeri and Capsicum annuum were moved to Group 2. Moreover, Cylindropuntia cholla and Celtis reticulata were included in the priority groups and Lysiloma microphyllum was excluded.

Discussion

Seeds as cereals and legumes are the staple foods for humanity and they have played a crucial role in human society development (^{Cordain 1999}); almond-like seeds are also important complements and source of proteins and lipophilic nutrients, thus a good Food Significance Index should be able to identify such kind of seeds as important complementary food. Fruits and vegetables play a similar role in diet; both are complements to the staple foods and are the main source of dietary fiber, vitamins and minerals. In our study, fruits showed higher FSI values than vegetables. People usually prefer fruits and their transformation into non-perishable or marketable products (e.g. jams, marmalades, dehydrated fruits, juices) is easier. Other authors have also reported higher preference for fruits than for vegetables (^{do Nascimento et al. 2013}, ^{Kujawska & Łuczaj 2015}).

The FSI and the SI showed moderate differences in the ordered rank of the species. SI was better than the FSI to reflect the preference (pr); thus, it was biased towards most popular species, but not always for the best options for food security. For example, seeds showed the highest FSI, but the fruits with the highest citation index (ci) were for SI. Seeds are more caloric and non-perishable foods and they could be better for food security; this failure was corrected in the FSI by including the Crop Value (CrV). Moreover, SI values for Malpighia diversifolia and C. reticulata (0.201–0.287) were similar to those of Cylindropuntia cholla and Cnidoscolus maculatus (0.294–0.318). The first two are very small fruits consumed only in the field as a candy and they hardly could deal with hunger; this failure was corrected in the FSI by including the food category (ca).

The FSI included higher culinary diversity in the higher valuated species than the SI. The first 12 species chosen by the FSI included four fruits, three seeds, two vegetables, two seasoning, and one tea. By contrast, SI only included seven fruits, three vegetables, one seed, and one tea. However, the selection of the most important species (Priority Groups) showed only minor differences among FSI and SI if it was included the culinary diversity system (selection by type of food), instead of a selection using only the rank values of both indices. For example, Pithecellobium dulce and Matelea cordifolia were the vegetables with the highest FSI values and the only two vegetables with a processed product; however, their FSI were lower than the most popular seeds, fruits and seasonings reported in this study. The exclusion of these species of the priority groups means a lost in the culinary diversity, and they would be excluded based only in the highest values of the index as suggested in previous researches (^{Joshi et al. 2015}); for instance, selection of only those with FSI values higher than 3.0. Alike, all seeds were excluded from Priority Group 1 using the SI values. Although, SI is easier to assess, it is recommended FSI because provides more information about the food (e.g. role in diet, postharvest shelf life, processed products) and not only its popularity as measured by the SI.

Different results are expected with others indices reported in literature. The Pieroni Food Significance Index (^{Pieroni 2001}) favors the wild green vegetables over other parts of the plants; however, we consider that other factors as the role in the diet or the postharvest shelf life are more important than just the organ that is consumed. For example, a representative green vegetable such spinach or lettuce cannot be more important for food security than those typical seeds used as staple foods (e.g. maize, rice, beans); moreover, two different seeds can represent different role in diet and so their importance should vary from a staple food such as chickpea to a complement such as peanut.

Unlike the reported for some indigenous communities (^{Eyssartier et al. 2011}, ^{Cruz et al. 2013}, ^{Sujarwo et al. 2014}, ^{Narchi et al. 2015}), knowledge erosion was no observed among the younger and the older male ranchers, at least in the number of species they cited. This result agrees with the reported for a non-indigenous community in Rayones, Nuevo León, Mexico; however, unlike our finding, they found that women cited more species (edible and no edible) than men (^{Estrada-Castillon et al. 2014}). A review from European communities shows economic activities such as farming and those involving long walks in nature are closely related with high traditional knowledge in the use of wild edible plants (^{Łuczaj et al. 2012}); such activities are very similar to the life style of the Sudcalifornian ranchers and are practiced from an early age, this might explain why younger and older men cited a similar number of plant species.

An outstanding feature of the ranchers was the importance of eating the seeds of oaks (Quercus spp.), which does not seem to be important for many indigenous cultures of Mexico (^{Camou-Guerrero et al. 2008}, ^{Vázquez-Alonso et al. 2014}). This custom was like the registered for the rural populations from Spain or Europe (^{García et al. 2014}). Acorns were an important food for pre Columbian people from Mesoamerica and California, but today they are considered as secondary or emergency foods (^{Clarke 1997}, ^{Mapes & Basurto 2016}, ^{Zizumbo-Villareal et al. 2016}). The development of this knowledge could be due to the Spanish origin of the missions combined with the hard conditions for agriculture establishment and the scarcity of basic grains. Therefore, they learned about the local source of seeds.

Recommendations for use, conservation, and domestication. Quercus brandegeei, Ficus petiolaris, Cyrtocarpa edulis, and Pithecellobium dulce (species in Priority Group 1) are slow-growing trees, and domestication process would be hard and slow. However, considering their food importance several strategies could be established such as: inclusion in local reforestation programs, monitoring the health of trees, determining the best harvesting season, and assessing their productivity. It must be also considered these three species are important for local wildlife. The application of these strategies is particularly important for Q. brandegeei, which is endemic to the Cape Region and abundant in the studied area. On the other hand, use and domestication of Cnidoscolus maculatus (placed in Priority Group 2) is more feasible because is a perennial herb, endemic and widespread in Baja California Sur state; thus, it could provide greater agronomic facilities and less ecological risk that a slow-growing tree. A similar case is for the vegetables such as Matelea cordifolia (Priority Group 1). Moreover, there are some progress about the management and domestication of several species of the genus Stenocereus and other Mexican columnar cacti (^{Casas et al. 1999}, ^{Parra et al. 2012}, ^{Pérez-Negrón et al. 2014}, ^{Pérez-González et al. 2015}), knowledge that must be the base of integral programs for the sustainable use of S thurberi and S. gummosus. By contrast, there are many regional studies about the domestication of Turnera diffusa and Capsicum annuum (^{Araiza-Lizarde et al. 2011}) that can be adopted for this area.

There are not nutritional studies about the acorns of Quercus brandegeei, the fruits of Ficus petiolaris, Stenocereus thurberi, Stenocereus gummosus, and Cyrtocarpa edulis, and the vegetable Matelea cordifolia. However, acorns have been reported as moderate source of protein (7.5 mg/100 g) and tocopherols (2–11.2 mg/100 g) (^{Cantos et al. 2003}, ^{Kilic et al. 2010}), whereas domesticated species of Ficus (F. carica, the common fig) is a good source of dietary fiber and calcium (^{Morton 1897}). In addition, only one preliminary study shows the nutritional composition of the Cnidoscolus maculatus seeds (misreported as C. angustidens), these are high in proteins (30–36 %) and lipids (26 %) (^{León-de la Luz et al. 1999}). By contrast, the Pithecellobium dulce arils have been better studied; they are rich source of dietary fiber, vitamin C and phenolic compounds (^{Pío-León et al. 2013}).

Accordingly with our discussion and considering their food importance in the studied area, their possible agronomic advantages, and the lack of studies in domestication or nutritional composition, future studies are recommended for the selected priority species and specially for Cnidoscolus maculatus, Stenocereus thurberi, S. gummosus, and Matelea cordifolia. Stenocereus thurberi and M. cordofolia which are widely distributed in the northwestern of Mexico; therefore, any progress in their domestication process might have local and regional impact. Others authors had documented the food importance of S. thurberi, and M. cordifolia for indigenous communities in the northwestern Mexico (^{Felger & Moser 1976}, ^{Yetman & Van Devender 2002}).

Conclusions

The Southern Baja California ranchers showed a good knowledge about the use of wild edible plants, and this was homogenously distributed in 20 to 81 years old men. Although Food Significance Index (FSI) and Salience Index (SI) showed differences, the plant species grouped in the Priority Group 1 were mostly the same when inclusion was decided considering information about culinary diversity. Thus, application of this strategy can improve the results of other ethnobotanical indices. Based on their food importance and potential agronomical facilities, Stenocereus thurberi, S. gummosus, Matelea cordifolia, and Cnidoscolus maculatus are the priority species recommended for further studies about composition, biological activities and domestication with the prospective of their future introduction as new crops for the northwestern of Mexico.