Ethnobotanical works carried out with wild food plants of Brazil without a high risk of bias were concentrated in two biomes: the Atlantic Forest and Caatinga. Only one study was performed in Pantanal. The predominance of the Atlantic Forest and Caatinga as study areas of works without a high risk of bias was also observed in a national-scale meta-analysis of patterns of use of medicinal plants [3]. Several factors may be underlying this trend, including issues related to the distribution and profile of research groups. A high number of ethnobiological studies have been conducted in ecosystems of Cerrado (savannas and seasonally dry forests), for example, but most of them have a qualitative nature and make use of theoretical samples that make sense for their objectives, but which are not suitable for meta-analyses.

Therefore, he absence of studies in the Amazon and Cerrado does not necessarily imply that these biomes are inadequately represented in terms of ethnobiological research efforts. It simply signifies that the epistemological orientation of these studies does not align with the criteria for inclusion in a meta-analysis. While we did not incorporate Amazonian studies into our research, there is evidence to suggest that, in certain aspects, their patterns of wild food plant consumption are not significantly different from those in other Brazilian biomes. For instance, a study conducted in the Amazon, which did not meet our inclusion criteria, gathered information from fieldwork in various communities and concluded that the consumption of greens in the region is low, as people tend to prefer wild fruits and tubers [41].

Regarding the temporal distribution of the studies, there was a slight increase in the number of studies without a high risk of bias starting in the year 2013. This may be related to the increasing number of courses and manuals about methods in this field of research that led to greater care in the selection of more robust samples. Furthermore, the studies of Kinupp & Barros in 2007 and 2008 [42, 43] and the further publication of the book “Plantas Alimentícias Não Convencionais” (unconventional food plants) in 2014 by Lorenzi and Kinupp [44] brought popularity to this topic and could also have helped to increase the number of research groups interest in studying unconventional or wild food plants.

Eliminating sampling problems from quantitative ethnobiological studies is fundamental from the point of view of biodiversity conservation because information with sampling problems and biased results may have an influence on decisions and actions. Biased results include, for example, misleading clues about the conservation status of some plants, the identity of the most popular plant species and the strategies necessary for management [16].

The elimination of study bias can be achieved through a good sample design, which means having a sample that accurately reflects the entire population while respecting the margin of error and confidence interval. However, not only does this ensure representativeness, but it is also necessary to adhere to the principles of randomness, thereby avoiding the influence of specific groups, particularly in the context of quantitative research [16]. It is also important to identify the plant material, specify that the material was identified by comparing voucher specimens or consulting experts, and provide a complete list of species [3].

Our findings point to a high predominance of use of fruits of food plants native to Brazil, which in turn influences the predominance of use of reproductive, non-persistent, and non-destructive parts. The predominance of fruits was also found in other studies at different scales, such as in the Yi peoples in China, where fruits were the most used part, followed by roots and shoots [45], and in the Kaski district (Nepal), where fruits were also the most used part, followed by young shoots [46].

However, leaves are the most consumed parts of wild food plants in different regions of the world, such as among the Vasavas in India [47], the Mapuche in Argentina [48], among ethnic minorities in Yunnan, China [9] and in two valleys of the Qinling mountains, Shaanxi, China [49].

The reasons for regional or national differences in use patterns may be plenty. First, it may have to do with the availability of edible parts in each region. For example, tropical ecosystems are more likely to produce fleshy fruits [50]. In many contexts, fruits may be preferred over green vegetables, considering that leaves are more likely to have bitter tastes, which are usually more avoided due to evolutionary processes (most toxic products in nature have bitter teste) [51]. Therefore, in contexts where food is not scare (e.g., high presence of fruits and other plant parts other than leaves), people would avoid using green parts of plants, in a process that is sometimes called herbophobia [52].

Moreover, in some tropical and subtropical regions there is a higher occurrence of thick leathery leaves [49], which are often not considered the most appropriate leaves for edible purposes. Those ecological patterns probably influence a higher use of fruits when compared to leaves in the tropics, but the existence of such prevalence needs to be further investigated.

On the other hand, cultural aspects may also play a significant role in shaping the consumption of leaves from wild food plants. For instance, in many local Chinese communities, the consumption of green leaves is more substantial compared to many other regions worldwide, making them some of the most herbophilous communities globally [52, 53]. The cultural forces that drive this high consumption of greens can be observed even in their language, as many plant species have popular names containing the word ‘cai’ (meaning ‘vegetable’), which encodes the edible nature of these greens in the language [53]. Furthermore, the cultural importance of greens is evident in the fact that these communities continue to practice traditions that have been lost in many parts of the world, such as the drying of wild vegetables for winter storage [53]

The patterns of preference for specific plant parts differ between food and medicinal plants in the Brazilian context: while fruits are preferred in the first group, leaves prevail in the second [3]. The preference for fruits for food purposes may be associated with the fact that this part has constituents with greater nutritional quality for consumers. In turn, leaves have a higher concentration of therapeutic agents, what explains why they are more frequently used in medicinal preparations.

Among the plant parts used for food, the parts classified as reproductive, non-persistent, non-destructive and parts of woody plants prevailed. However, the type of ecosystem had no influence on the use patterns, that is, local populations in seasonally dry and moist environments did not differ in terms of use patterns. Thus, it is possible that people’s preferences are not influenced by seasonality, but by other environmental conditions and factors. In the case of persistence and habit, the present study did not support the seasonality hypothesis, contrary to works focused on medicinal plants [3, 15].

An important factor that explains these differences is that although persistent parts of plants are important sources of compounds with medicinal properties, they are often not suitable for consumption as food, given that a small number of species have tubers or similar parts, or even stems with food potential.

The fact that wild food plants were not the food base of the groups in most studies may also explain the absence of strategies to secure the access to these resources in seasonally dry environments. When other products meet dietary needs, the spatial or temporal availability of wild food plants may lose relevance in relation to other variables, such as flavor [54].

Thus, considering the complementary role of wild food plants in the diet of most Brazilian populations and from the point of view of the socio-ecological theory of maximization, even if fruits and other non-persistent parts are not available year round, other variables may confer a great advantage to these resources to the point that availability becomes secondary [55].

From the point of view of conservation, the predominance of the use of non-destructive plant parts, especially fruits, places food use among factors with a low potential impact, corroborating the literature regarding non-timber forest products [12]. Such lower potential impact favors the stimulation of sustainable use and even the popularization of wild food plants as a strategy for food diversification and for increasing food and nutritional security. However, it is necessary to consider that, despite the lower impact in comparison with, for example, timber products, the use of wild food plants requires management strategies aimed at the preservation of plant populations. In this sense, the predominance of use of reproductive plant parts indicates the need for strategies to monitor the recruitment of new individuals, which can be operationalized through participatory management (or co-management). Participatory management is understood as the cooperation between government agencies, traditional local communities, resource users, as well as non-governmental organizations and other stakeholders, sharing the management and responsibility for an area or a set of resources [56]. In the context of native wild food plant use in Brazil, we recommend training local harvesters to conduct both qualitative and quantitative monitoring of seedlings. Their constant presence in the harvesting areas will enable more frequent monitoring. Additionally, the entire process should not be top-down planned, as communities should also have a say in determining the main purposes and objectives of the conservation strategies.