Physicochemical composition of Tamarindus indica L. (Tamarind) in the agro‐ecological zones of Uganda

Abstract The relationships between the physicochemical composition of Tamarindus indica pulp and seeds, and agro‐ecological zones and land use types were assessed in Uganda. The objective was to determine the relationship between the physicochemical composition, agro‐ecological zones, and land use types. The samples were processed by manually depulping the T. indica pods, sun‐drying the pulp and seeds, and grinding into powder. The powdered samples were analyzed for β‐carotenoids, vitamin C (ascorbic acid), calorific value, crude oil, acid, and peroxide values. Data were analyzed using ANOVA in the general linear model (GLM). Principal component analysis (PCA) was used to relate the physicochemical properties to the agro‐ecological zones and land use types. There were significant differences (p ≤ .05) in the physicochemical composition variables between agro‐ecological zones and land use types. Land use types showed strong correlations with physicochemical properties while agro‐ecological zones did not show correlations. The results show that in terms of general properties, T. indica pods provide a valuable, rich, and exceptional source of vitamin C, compared to many widely consumed indigenous and conventional fruits and vegetables. The pods from land use types characterized by natural habitats had relatively more nutrient levels than the land use types influenced by anthropogenic activities.

Many studies have determined the physicochemical composition of common fruits and vegetables (Ajayi, 2010;MFAF, 2014;Singh et al., 2007;Tagoe, Dickinson, & Apetorgbor, 2012). However, in Uganda, T. indica, an indigenous species has been neglected, indiscriminately grows without much attention at anywhere like on-farm, wild, roadsides, abandoned homesteads, market places but are cheap and are commonly consumed by rural population. The use of T. indica pulp, leaves, and flowers for food, medicine, and other industrial purposes (e.g., juice, concentrate, powder, pickles, and paste) has been reported by Singh et al. (2007). Okello et al. (2017) reported that the T. indica fruits provide two important products-pulp, mostly eaten directly or used for making local food, drinks, and sold for domestic incomes, whereas the seeds, obtained after depulping the pod, are usually thrown away due to total lack of knowledge on its properties and uses.
Whereas there is information on the physicochemical properties of many neglected indigenous tree species (Abiodun, Dauda, Adebisi, & Alonge, 2017;Ajayi, 2010;Hiremath, Yadav, & Suguna, 2016;Ouilly et al., 2017;Sulieman et al., 2015), many studies focused on the edible parts (pod pulp and leaves). Little attention was given to non-edible parts such as seeds. Furthermore, the relationship between the physicochemical characteristics and growth conditions within the different agro-ecological zones and land use types has not been investigated despite reported uses. Thus, the objective of the present study was to determine the T. indica physicochemical composition of T. indica in relation to agro-ecological zones and land use types in Uganda. It was hypothesized that there were no significant differences in the physicochemical characteristics of T. indica pulp and seeds between agro-ecological zones and land use types. The F I G U R E 1 Map of Uganda showing agro-ecological zones and sample sites findings of this study are useful for guiding the utilization and land use-related decisions.

| Agro-ecological zones of Uganda and study sites
Uganda, officially the Republic of Uganda, is located astride the Equator in Eastern Africa covering 2,41,550.7 km 2 . It lies between latitude 4°12′N and 1°29′S and longitude 29°34′E and 35°0′E. Most of Uganda lies within the interior plateau of the Africa continent. The predominant rocks were formed between 3,000 and 6,000 million years ago (pre-Cambrian era). However, in some western and eastern parts of the country, there are major developments of younger rocks, which are either sediments or volcanic in origin ranging from 135 million years ago (Cretaceous period) to the present day (NEMA, 2007). The types and characteristics of the soils of Uganda are defined by parameters such as nature of the parent rock, age of the form, and climate (especially the amount of moisture). The ferralitic soils (ferralsols) are the most dominant. Others are ferruginous, volcanic, and alluvial soils (NEMA, 2007). The climate is tropical, with distinct wet and dry seasons. In many parts of the country, the dry season occurs between December and March.
Uganda is divided into nine agro-ecological zones primarily based on agro-climatic factors (rainfall totals and distribution) and soils (productivity and fertility). Topography, temperature, moisture, and vegetation cover are the secondary factors characterizing a zone but differing between zones (GoU, 2004). Hence, the climate, geological formation, topography, soil types, rainfall, and farming systems or practices are fairly homogeneous within zones. This study was conducted in the three agro-ecological zones (Eastern, West Nile, and Lake Victoria Crescent) where T. indica trees occur naturally on-farm and in the wild. Each zone was represented by a district, that is, Soroti, Moyo, and Nakasongola for Eastern, West Nile, and Lake Victoria Crescent agro-ecological zones, respectively ( Figure 1). Uganda is divided into 121 districts (including the capital city of Kampala), which are grouped into four administrative regions. The district is the largest decentralized unit of administration in the country. Nakasongola district is located in Central Uganda in the Lake Victoria Crescent agro-ecological zone. It covers 3,510 km 2 , and the altitude ranges between 1,000 and 1,400 masl. The topography is generally flat, characterized by small altitudinal differences with poor drainage in the wide flat valleys and Lake shores. It is endowed with unique rocky outcrops (isenbergs). The soils are mostly shallow and skeletal having developed from quartzite or iron stone (ferralitic). The vegetation type is open deciduous savanna woodland.
There are two rainy seasons (March-July and August-November), with bimodal total rainfall of between 875 and 1,000 mm per annum.
Soroti district is located in Eastern Uganda in the Eastern agroecological zone, covering an area of 2,662.5 km 2 and altitude of 1,036-1,127 masl. The major rocks are granites, magnalites, gneiss, schists, and quartzites. The district is mainly underlain by rocks of the basement complex pre-Cambrian age. The major soils are of the Serere and Amuria catena, Metu complex, and Usuk series with moderate agricultural productivity. The vegetation types include wooded savanna, grassland savanna, forest, and riparian. Two rainy seasons occur from March to June and August to November with total annual rainfall of 1,000-1,500 mm, with December and January being the driest months. The minimum and maximum temperatures are 18 and 31. 3°C, respectively (GoU, 2014).
Moyo district is located in the West Nile region of Uganda within the West Nile agro-ecological zone. It covers an area of 1,891 km 2 with an altitudinal range of 600-1,586 masl. The topography is characterized by low plains as well as rolling hills and valleys that slope toward River Nile. A series of hills and peaks characterize the northern and northeastern parts of the district (NEMA, 2004b). The major geological formations are the gneiss, alluvial deposits, schists, quartzite, and marble that occur in the mountains. The major soil types include the vertisols, leptosols, alluvial deposits, and ferralsols that are moderately fertile (NEMA, 2004b). The rainfall varies between 1,500 and 1,700 mm, less pronounced, bimodal, and occurring mainly in March to June and August to November with the dry season in late November to early March. The minimum and maximum temperatures (23.7-30°C) are of modified equatorial type (UDIH, 2007). The vegetation is classified as wooded savanna.

| Sampling design
The sampling sites were located within an ecological gradient that took account of differences in climate and agro-ecology. The three sample agro-ecological zones (Eastern, West Nile, and Lake Victoria Crescent) are located more than 300 km apart. Each agro-ecological zone was stratified into two major land use types: crop fields (onfarm) and wild lands. The crop fields (on-farm) are farmlands with agricultural crops, while the wild land had not been cultivated for 5 or more years prior to the study.
Four representative sub-counties (sample sites) were selected in each land use type in each district making a total of twelve study sites (four study sites per district) covering about 5 km 2 . Up to five T. indica sample trees were randomly selected per land use type in each sampling site based on ease of access, absence of obvious signs of pests, diseases and fire, and presence of good mature pods. The sample trees were located at least 200 meters apart to avoid sampling siblings.

| Sample collection methods
The pods were collected during the dry season between December and March. Fruits were collected from the top, middle, and bottom parts of the tree canopy. The trees were climbed using ladders. Ripe pods were selected by gently squeezing them. The ripe pods had scurfy brown, woody, fragile shells with brown pulp that cracked on squeezing. The stalk of the ripe pod was severed with a knife to remove the pod without causing damage to the pod, developing flowers and leaves.
Eight pods were collected from each canopy level making 24 pods from each tree. A total of 480 pods were harvested from 20 trees per land use type, giving 960 pods per district (zone) and 2,880 pods from the three agro-ecological zones. Pods collected from each tree were pooled, kept in white polythene bags and labeled. The samples were then taken to Makerere University's College of Agricultural and Environmental Sciences laboratory for physicochemical analyses.

| Samples preparation for laboratory analyses
The pods were washed with distilled water and allowed to dry for 1 hr to maintain constant moisture content of the shell. Pod shells were manually depulped, and morphological traits such as pod length, pod breadth, pod total mass, pod seed number, total seed mass, and pulp mass were determined. All measured samples were pooled by land use types and agro-ecological zones for proximate analyses.
Decomposed and damaged pulp and seeds were discarded. The depulped seeds and pulp were separately sun-dried for 6 hr/day for 3 days to lower the moisture content and later dried in an oven at 40°C for 3 days to 8% moisture content. These were then separately grounded in an electric grinding machine (Brooks Crompton, 2000 series-UK) to 60-mesh size. The powdered samples were stored in ziplock plastic bags at room temperature for further laboratory analyses of β-carotenoids, vitamin C (ascorbic acid), crude oil, acid (gmg -1 ), and peroxide (mEq/kg), whose values are very important for human and animal nutrition.

| Crude oil extraction and physicochemical analyses
The powder from pulp and seeds was separately analyzed for β-carotenoids and vitamin C (ascorbic acid). The β-carotenoids were analyzed under subdued light following Rodriguez-Amaya and Kimura (2004), while vitamin C was determined by titration based on AOAC (1999). Oil was first extracted using the soxhlet methods (AOAC, 1999) and then subjected to physical and chemical characterization. The acid value (gmg -1 ) and peroxide value (mEq/kg) were determined using the analysis of rancidity (Kirk & Sawyer, 1989). All physicochemical composition analyses were performed in triplicate and the average reading recorded.

| Physicochemical composition of T. indica pulp samples
Generally, there were differences between agro-ecological zones and between land use types. The physicochemical composition values of the pulp were higher in the West Nile agro-ecological zone samples as well as the samples collected from the wild land use types. Specifically, there were significant differences (p ≤ .05) between agro-ecological zones and between land use types for βcarotenoids as well as vitamin C. The samples from the West Nile zone had significantly higher vitamin C contents compared to samples from other agro-ecological zones. However, β-carotenoids were nearly similar in all the agro-ecological zone samples (Table 1).

| Physicochemical composition of T. indica seed samples
The physicochemical composition of T. indica seed samples showed significant differences (p ≤ .05) between agro-ecological zones. All properties were significantly different between land use types with the exception of peroxide value. The absolute amounts of different components were nearly similar in all agro-ecological zones and land use types (Table 2).

| Relating the physicochemical properties of pulp samples to agro-ecological zones and land use types
The first principle component (PC1) is strongly correlated with all the two original variables (β-carotenoids and vitamin C). It also shows a perfect contrast between these variables (β-carotenoids and vitamin C). The first principal component increases with increasing βcarotenoids and decreasing vitamin C contents (Table 3)  The contents are expressed by the mean values ± SD for n = 3 and n = 2. AEZ, Agro-ecological zone. *Interactions between land use types and agro-ecological zones with variables; same superscript letters within a row show no significant difference.
Accordingly, this component can also be viewed as a measure of how poor contents of vitamin C affects its functions in humans (prevention and treatment of scurvy, an antioxidant and co-factor functions for enzyme metabolism), and plant (co-factor for enzymes involved in photosynthesis and synthesis of plant hormones).
In the second principal component (PC2)

| Relating the physicochemical properties of seed samples to agro-ecological zones and land use types
The first principal component (PC1) is strongly correlated with three of the original variables (β-carotenoids, calorific and acid values).
The component also has large positive and negative associations with these three variables as shown in

| Vitamin C (ascorbic acid)
The  (2017) shows the kiwifruit tree species as an exceptionally rich source of vitamin C but a kiwifruit that has been cool stored for a while has reduced vitamin C content.
Additionally, Liji (2016) reported that the availability of vitamin C is reduced by cooking and long periods of storage.
The differences observed in terms of the contents of vitamin According to Muhammad et al. (2014), vitamin C levels in the unripe fruits are higher than the ripe ones but generally decreased upon increase in temperature, ripening, and time of exposure. While species such as Ziziphus jujube fruit, the vitamin C content does the opposite-it rises with increased ripeness (The Natural Food Hub, 2017). Kaleem et al. (2016) also reported that at increased temperature, the amount of vitamin C decreases, due to oxygen, the most destructive element in fruit juice causing degradation of vitamin C in fruits. At increased temperature, the juice is more susceptible to oxidation and the effects of temperature and consequently experience more degradation of vitamin

| Beta carotenoids (β-carotenoids)
The amount of β-carotenoids in T. indica pulp and seed samples were 0.14-0.17 mg/100 g and 0.13-0.33 mg/100 g, respectively, in this study. The most important of these are α-and β-carotene, which are precursors of vitamin A. Of the carotenes, β-carotenoid has the strongest provitamin A activity. Carotenoids are important in human health. Vitamin A deficiency can lead to blindness, and the use of

| Acid value
Acid and peroxide indexes are parameters that demonstrate the quality of the oil (Muniz et al., 2015). Acid value of 10.0-19.5/gmg was recorded for this study. Akubugwo, Chinyere, and Ugbogu (2008) reported that the acid value indicates edibility of oil and its suitability for use in the paint industry. Fats and oils are graded by their acid and free fatty acid contents, which are used as an index to determine their quality (Kardash & Tur'yan, 2005). The values of the present study are higher than those documented for Lannea kerstingii (Ouilly et al., 2017) and Allanblackia floribunda (Wilfred, Adubofuor, & Oldham, 2010). However, our documented values are within the range reported by Taha, Nour, and Elkhalifa (2016) but lower than those values reported in other studies such as Ajayi (2010) and El-Siddig et al. (2006). However, a study by Muhammad, Ayub, and Zeb (2013) documented low acid values for olive fruit (Olea europaea), while Kittigowittana, Wongsakul, Krisdaphong, Jimtaisong, and Saewan (2013) reported higher values for Hervea brasiliensis.
Oils processed from fresh fruits have low free fatty acids (low acid values) compared with those from many days' old fruits. The antioxidant activities of the seed oils may be attributed to fatty acid components (Kittigowittana et al., 2013). Fatty acids, which are the major constituents of oils-may include saturated, monounsaturated, and polyunsaturated fatty acids which contribute to human physiology in different ways (Costa, Ballus, Teixeira, & Godoy, 2011;). such, an increase in saturated fatty acid content is specific in seeds and fruits. Additionally, highly unsaturated vegetable oils are less suitable for many food applications (Satchithanandam et al., 2004).
A high proportion of the variation in fatty acids content is due to environmental factors (Lanna, José, Oliveira, Barros, & Moreira, 2005). Our study reported higher acid values in West Nile agroecological zone with maximum temperature above 30°C than Lake Victoria Crescent agro-ecological zone with maximum temperature below 30°C. All fats and oils in nature are a mixture of saturated, monounsaturated and polyunsaturated fatty acids -the difference in them is the proportion of each. Whether in plant or animal tissues, fats and oils operate at different temperatures, the most important consideration which is often neglected when discussing their healthiness. The degree of saturation or unsaturation determines not only a fat's melting point, but also its chemical stability and its likelihood of auto-oxidizing and creating harmful free radicals. The degree of saturation of plant oils and fats is entirely dependent on the temperature in which they are grown. The higher the proportion of saturated fatty acids a fat is, the less likely it is to go rancid; the more polyunsaturated fatty acids it contains, the more difficult it is to stop it going bad (Barry, 2011).
Fruits that have been stored for long periods prior to processing and determination have high free fatty acid contents. Fatty acids play a very important role in fats and oils because of their health implications in the human diet and properties in industrial processes.
In addition, Ekop, Etuk, and Eddy (2007) reported an increase in the acid value of oil is an indication of the onset of rancidity. However, hydrogenated fats and oils prevent rancidity and are used in foods to improve texture and stability for a longer shelf life because trans fatty acids have higher melting points and greater stability than their cis isomers (Dixit & Das, 2012). And rancidity can be accelerated by moisture, air, and presence of some metals, which agrees with a report by Okello et al. (2017; who documented high levels of moisture content and mineral composition in the Lake Victoria Crescent agro-ecological zone, which recorded high acid value.
Oil extraction and acid value determinations for our study were delayed, and this may explain the relatively high acid values recorded.
Our study findings corroborate the findings of Tagoe et al. (2012) showing that free fatty acid content of tamarind oil was affected by the length of sample storage and the length of storage of the oil after processing. According to Tagoe et al. (2012)

| Peroxide value
The peroxide values of T. indica seed oil samples ranges from 111.0 to 235.1 mEq/kg. Peroxide value indicates the degree of oxidation of the oil-low peroxide value indicates low oxidation (Ekop et al., 2007).
The present study values are above the values reported for most common plants (Ajayi, 2010;Ekop et al., 2007;Muhammad et al., 2013;Ouilly et al., 2017;Wilfred et al., 2010). This value is above the range of peroxide values (<10 mEq/kg) generally reported for fresh fats and oils (Kirk & Sawyer, 1991 This is of no nutritional interest to humans, suggesting that they could not be utilized as edible oils which allows it not to be consumed as virgin edible oil. These antioxidants act as scavenger for damaging oxygen free radicals. This therefore indicates the ability of such oils to resist lipolytic hydrolysis and oxidative deterioration.

| Calorific value
The  (2002), the recommended mean energy intake for both male and female population of age group 18-30 years is 2,800 kcal/day but lower in males and females of 60 years and above. Our study found out that the plant species can contribute to the calorific requirement of the body and its consumption will contribute to the calorific requirements of the body. In order to ensure a calorie-controlled diet, regular consumption of T. indica is recommended.

| CON CLUS ION
There were significant differences in the physicochemical composition variables between agro-ecological zones and land use types.
The land use types showed strong correlations with physicochemical properties, while agro-ecological zones did not show correlations. The T. indica pulp contains higher vitamin C (ascorbic acid) than the seed samples. The seed samples, however, contain slightly more amounts of β-carotenoids than pulp samples. In addition, West Nile agro-ecological zone and wild land use type recorded superior physicochemical contents than other agro-ecological zones and land use types. The contents of vitamin C in this study show that the species is a valuable, rich, and exceptional source of vitamin C than most known indigenous and conventional fruits and vegetables. Its deficiency diseases in human such as scurvy and pellagra can be prevented through consumption of T. indica pulp and seeds. It is also important to consider improving fatty acid composition of dietary T. indica oils to enhance health and safety. The documented results thus offer scientific basis for use of T. indica pulp and seeds, especially the latter which are always discarded but are nutritionally important for human and animal diets. With a deliberate nutrition education efforts on the benefits of the species, the wild population that showed superior qualities than on-farm species requires more protection from anthropogenic activities that lead to the loss of these population.

ACK N OWLED G M ENTS
The authors are most grateful for financial support from Norwegian Agency for Development Cooperation (NORAD) through Makerere University. We are also indebted to the field assistants, laboratory staff, and the farmers whose fields were visited during data collection and that have made the compilation in this manuscript possible.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interests.