Journal of Society & Technology 5:15-24 (2015)

Phenotypic Variation in Tree Form Dimensions, Fruitand Seed Traits, and Oil Content of Jatropha(Jatropha curcas L.) from Three Provenances

Shierel F. Vallesteros1* and Wilfredo M. Carandang21Nueva Vizcaya State University, Bayombong, Nueva Vizcaya, Philippines2Institute of Renewable and Natural Resources, UPLB, College Laguna, Philippines

Abstract

Heightened attention to Jatropha (Jatropha curcas L.) was driven by the belief that it can bemade to yield the desired quantity of product in marginal lands, and growing it in marginallands would prevent the food and fuel competition. Many projects on seedling productionand plantation establishment were undertaken. However, identification of provenances thatgives higher seed yield and oil content is deemed necessary before any massive plantationestablishment ventures. Three provenances were selected, namely: Bacolod, South Cotabato,and Talisay. Regarding growth performance or tree dimensions, they all performed similarly.Variability was large in all morphological traits, the highest being in length of unbranchedmain stem (CV = 46.51%) and a number of first order branches (CV = 33.72%). Seed yieldwas positively and significantly correlated with basal diameter, length of unbranched mainstem, and crown diameter. Variability was small in fruit and seed traits. Crown diameterappeared to be a predictor (p<0.05) of three seed size parameters, namely: length, breadthand thickness. Among the seed traits, seed length was significantly correlated (r=0.498) withoil content. The oil content in kernel ranged from 41.40% to 59.26%..

Keywords: Jatropha curcas L., ideotype, biodiesel, biofuel

Introduction

A few years ago, the plant named jatropha(Jatropha curcas L.) or Tubang Bakod orTuba-tuba had become an instant celebrity inthe Philippines for being considered as themost promising biofuel crop. The promisesinclude the natural ability of the plant to growin poor sites or marginal lands, therefore,averting the dreaded food and fuelcompetition (Rajagopal, 2007; Kheira, 2009).Jatropha is also adapted to a wide range ofclimatic and soil conditions, as it grows inmany areas around the world, and may beused to improve the land that may lead to theintegration of food crops into plantations andfarms. The government and the private sectorrushed into finding lands for plantations,massive seedling propagation and engagingcommunities into small jatropha projects.Because of the excitement, they seemed toforget the need to do basic researches on the

plant, do site trials and explore differentprovenances or genetic resources from withinthe country.

While it is true that the species can easilybe propagated by seeds or cuttings andrequires simple maintenance activities, it isstill important to conduct researches on thespecies’ or variety’s growth and yield.According to Heller (1996), only little geneticresearch has been conducted so far forjatropha. Kaushik et al. (2007) added thatlittle work has been done so far on germplasmcollection and its evaluation for chemicalcomposition of seeds in India, one of theleading producers of jatropha oil. Moreover,Jongchaap et al. (2007) stressed that geneticidentification of provenances and testing themin different locations and conditions should bea priority research for jatropha. According toRajagopal (2007), the current trend injatropha cultivation is focused towards the use

*Correspondence: sfvallesteros@gmail.com

Journal of Society & Technology Vallesteros & Carandang

of improved germplasm for optimal yield ofseeds with an optimal amount of oil, andmatching this germplasm to sites withrelatively fertile soils and adequate moistureto enable it to reach its genetic potential toproduce optimum yields.

An initial research towards cropimprovement may include the development ofideotypes of jatropha. Ideotype refers to theconceptual model which explicitly describesplant phenotypic characteristics that arehypothesized to produce greater yield (Martinet al., 2001). Ideotype is mostly based onmorphological characteristics but can alsospecify biochemical, physiological, or otherattributes of desirable phenotypes. Most treeimprovement programs begin with selection ofpromising phenotypes from naturalpopulations (Martin et al., 2001) or fromdifferent provenances.

Wishing to contribute to basic researcheson jatropha, the study explored thephenotypic variations of three provenancesfrom the Philippines and looked into therelationships between several tree formdimensions, fruit and seed traits, and oilcontent of jatropha.

Materials and Methods

Location and climateof study area

The experimental jatropha plantation issituated in the Mt. Makiling Forest Reserve inLos Banos, Laguna. Its specific geographicallocation is 121.2362o east longitude and14.1627o north latitude. Figure 1 shows thelocation of the study relative to the fourjatropha provenances.

The plantation was established in theCollege of Forestry and Natural Resources ofthe University of the Philippines Los Banos. Itis part of the Biofuel Research and EnterpriseDevelopment Program being implemented by17 state universities and colleges in thecountry since 2006.

The site is approximately 40 m x 70 m gap

Figure 1: Location of study.

within a dipterocarp forest. The slope rangesfrom 1o to 4.5o while the elevation isapproximately 75 m above sea level. Thestudy site falls under Climatic Type I of theModified Coronas Classification. An averageannual rainfall based on 1977 to 2003 periodwas recorded to be 1,970 mm while the meanannual temperature was 27.5oC. The hottestmonths were April and May while the coldestmonths were January and December. Duringthe same period, the average solar radiationwas 4,358 langley.

Location and climateof provenances

Seeds from the different provenances used inthe study were provided by the PhilippineForest Corporation. The provenances werenamed after the municipality, city or provincethey were collected from. The jatropha seedsoriginated from the following: (a) Talisay,Batangas in Luzon (b) Bacolod City, NegrosOccidental in the Visayas, and (c) Tupi, SouthCotabato in Mindanao. Talisay falls under

16

Journal of Society & Technology Vallesteros & Carandang

Table 1: Geographic and climatic information on provenances.

Name of Provenance Latitude Longitude Rainfall Temperature(oN) (oE) (Mm) (oC, Min-Max)

Talisay (TL) 14.1041 121.017 1,950.80 23.2-31.9Bacolod (BC) 10.6475 122.981 2,642.30 22.1-32.8South Cotabato (SC) 6.3836 124.966 925.6 22.5-32.9

Climatic Type I, Bacolod City under ClimaticType III, and Tupi under Climatic Type IV ofthe Modified Coronas Classification.

The geographic coordinates and someimportant climatic measurements are shownin Table 1. The single-pair coordinateinformation for each provenance representsthe centroid of the municipality where theseeds came from. All climatic data wereobtained from the Philippine Atmospheric,Geophysical and Astronomical ServicesAdministration (PAGASA).

Experimental layout

The experiment was laid out in RandomizedComplete Block Design (RCBD) whereinsunlight was the factor that was addressed byblocking. A light meter was used to determinethe light intensity in each block. The lightintensities were 3,500 to 9,000 lux, 9,100 to14, 500 lux, and 14, 600 to 24,000 lux forBlocks III, II and I, respectively. Eachprovenance was replicated three times hencethree blocks were established. Each replicateor plot consisted of 25 plants, thus, the totalnumber of plants was 225. The dimension ofthe plot was 10 m x 10 m and the spacingbetween plants was 2 m x 2 m.

Dimensions measured

Several data were collected to describegrowth, fruit and seed traits, and kernel oilcontent. Tree form consists of basal diameter,crown diameter, length of unbranched mainstem, number of first order branches, andinternode diameter.

Basal diameter refers to the girth of themain trunk taken at 5 cm point above the

ground using a digital caliper (graduated to0.01 cm) when the plant is small or adiameter tape when the diameter exceededthe capacity of the caliper. Crown diameterwas taken by measuring two perpendicularlengths passing though the crown center. Itwas measured during the month when peakharvest was obtained (14 MAP). Length ofunbranched main stem was obtained bymeasuring the distance between root collarand the base of the first major branch using ameter stick. The number of first-orderbranches was counted 10 MAP. Internodediameter (in mm) is the diameter of the stemtaken just below the point where fruit stalkemerges.

Fruit and seed traits consist of fruit weight,fruit length, fruit breadth, fruit thickness,number of fruits per bunch, whole seedweight, seed length, and seed thickness. Fruitweight was taken within two hours aftercollection using the digital weighing scale.The length, breadth, and thickness of the fruitwere taken using digital caliper. The shorterdiameter of the fruit was assigned to breadthwhile the longer diameter to thickness.Weights were expressed in grams while lengthswere in millimeter. Measurement proceduresfor the seed followed that of the fruit. Thenumber of fruits per bunch refers to the totalnumber of fruits collected from each bunch,which is the cluster of fruits emerging from anode regardless of number of petioles.

Oil content pertains to percent crude fat inkernel, which was obtained using the laboratoryprocedure described at Table 2.

17

Journal of Society & Technology Vallesteros & Carandang

Table 2: Tree form dimension of the three provenances.

ProvenanceParameter

Height(cm)

BasalDiameter (cm)

Number of FirstOrder Branches

CrownDiameter (cm)

Talisay 264.46a 9.73a 3.20a 180.73a

Bacolod 265.53a 10.00a 4.37a 156.40a

South Cotabato 283.21a 10.10a 3.27a 179.57a

Grand Mean 264.97 9.28 3.4 160.76

*Means within a column followed by the same letter are not significantly different at 5% level byDuncan’s Multiple Range Test (DMRT)

Analysis of percent crude fat

After air-drying, the kernels were separatedfrom the seed coats, and then samples weredrawn and brought to the Central AnalyticalService Laboratory of the National Institute ofMolecular Biology and Biotechnology(BIOTECH) in Los Banos, Laguna foranalysis. Soxhlet Method as described by theAssociation of Official Analytical Chemists(AOAC 13th Edition, 1980) was employed.For each sample, two sub-samples wereprepared for the analysis. If the discrepancybetween the two sub-samples wasconsiderably high, the analysis was repeateduntil the discrepancy became within 0.01% to0.18% of the crude fat content.

Due to cost consideration, only 24 sampleswere randomly taken from a set of individualtree data for percent crude fat.

Results and Discussion

Tree form dimensions

It is the interest of this research to infer treemodel that may project yield and quality ofmain product which in a case of jatropha isthe oil content; hence, the growthperformance parameters were considered astree form dimensions.

Based on four growth parameters, the threeprovenances grew similarly in the experimentalarea 18 months after planting. This resultagrees with that of Sukarin et al. (1987 ascited in Heller, 1996) work which reported theabsence of morphological differences between

42 jatropha clones originating from differentlocations in Thailand. However, Heller (1996)reported significant differences in thevegetative development among 13provenances of jatropha in multi-location fieldtrials in Senegal and Cape Verde. The siteshave a semi-arid climate, with a short rainyseason (approximately four months) and alonger dry season of approximately eightmonths with a wide variation in rainfall(between 200 and 800 mm).

Correlation betweentree form dimensions

The relationships between tree formdimensions are summarized in Table 3. Thetallest trees were not necessarily the ones withthe biggest basal diameter, longestunbranched main stem, and highest numberof first order branches. Height was weaklycorrelated with the basal diameter (r=0.298),crown diameter (r=-0.111), length ofunbranched main stem (r=0.007), andnumber of first order branches (r=-0.117).The result suggests wide variation in tree formis expected even if trees are regularly spacedin plantation.

Basal diameter was significantly andpositively correlated with crown diameter(p<0.01), number of first order branches(p<0.05) and internode diameter (p<0.05).Internode diameter was also significantly andpositively correlated with height (p<0.05).

The coefficients of variation were large inall the morphological traits, the highest beingfor length of unbranched stem (46.51%) and

18

Journal of Society & Technology Vallesteros & Carandang

Table 3: Correlation coefficients between morphological traits.

Tree Form DimensionsTree Form Dimensions

HeightBasal

DiameterCrown

Diameter

Length ofUnbranchedMain Stem

Number ofFirst OrderBranches

Height 1Basal Diameter 0.298 1Crown Diameter -0.111 0.661 1Length of unbranched main stem 0.007 -0.35 -0.152 1Number of first order branches -0.117 0.4 0.298 -0.085 1Internode Diameter 0.551 0.432 0.15 0.064 -0.086

Note: Critical value of r at 5% level of significance, n=24, is 0.396, and 0.505 for 1% level (Snedecorand Cochran, 1967 as reprinted with permission by Gomez and Gomez, 1984)

number of first order branches (33.72%). Thelarge CV values likewise suggest that highvariability exists among individual trees. Whilethese morphological traits, especially height,basal diameter and crown diameter, may beeasily affected by site factors, it must benoted that trees were assumed to have beenexposed to similar site conditions especiallysunlight which was the basis of blocking.

It is interesting to note that basal diameteris negatively correlated (r=-0.350) withlength of unbranched main stem. It meansthat the large basal diameter was associatedwith increasing number of first order branchesnear the root collar. J. curcas has a shrubbyhabit (Heller, 1996; Jongschaap et al., 2007)and, therefore, the increase in basal diameter,which is an indicator of good health of theplant, is accompanied by increasing number ofbranches along the basal diameter point.Large-diameter trees will support morebranches, and higher yield is obtained in treeswith more branches (Villancio, 2008). Thesilvicultural implication of this relationship isthat spacing can be widened without resultingin lower yield. In addition, a smaller numberof trees could be grown in the plantationthereby reducing fertilizer and maintenancecosts.

The internode diameter ranged from 9.76mm to 16.29 mm, giving a mean of 13.59mm. It was weakly correlated with thenumber of fruits per bunch, which means thatthe size of the fruiting stem - the

measurement based on point close to thebunch - was not a determinant of the numberof fruits per bunch. It implies that fruits mayemerge from many points along a fruitingbranch.

Correlation between yieldparameters and tree formdimensions

Both basal diameter and length of unbranchedmain stem were positively correlated (p<0.05)with all yield parameters except with thenumber of fruits per bunch (Table 4). Crowndiameter, which was strongly and positivelycorrelated (p<0.01) with basal diameter, waslikewise positively correlated (p<0.05) withthe yield parameters except with fresh weightof seeds and number of fruits per bunch. Onthe other hand, the number of first orderbranches was positively correlated (p<0.05)with fruit weight and fresh weight only. Ingeneral, yield can be predicted using any ofthe growth performance parameters exceptheight and internode diameter. Among theyield parameters, however, basal diameter andlength of unbranched main stem were mucheasier to measure. Because of the very strongpositive correlation between most of the yieldparameters, tree breeders may use only one ofthese parameters. For farmer’s use, the choicebetween weighing (e.g., weighing of fruits)and counting (e.g., counting of fruits) shall beevaluated.

19

Journal of Society & Technology Vallesteros & Carandang

Table 4: Correlation coefficients (r) between tree form measurements and yield parameters.

Morphological TraitsYield Parameters

Weightof fruitsper tree

Freshweightof seeds

Number offruits per

tree

Number ofseeds per

tree

Number offruits/bunch

Basal diameter 0.401 0.491 0.452 0.452 0.146Height -0.036 0.022 -0.06 -0.06 0.198Crown diameter 0.401 0.368 0.475 0.475 0Length of unbranched main stem 0.428 0.399 0.439 0.439 0.203Number of first order branches 0.445 0.442 0.377 0.377 0.113Internode diameter 0.218 0.293 0.192 0.193 -0.086

Note: Critical value of r at 5% level of significance, n=24, is 0.396, and 0.505 for 1% level (Snedecorand Cochran, 1967 as reprinted with permission by Gomez and Gomez, 1984)

Height was not correlated to any yieldparameter. Moreover, the correlationcoefficients between height and fruit weight ornumber of fruits or total number of seedswere negative.

Fruits and Seed Traits

For a tree crop such as J. curcas, a plantmodel or ideotype often includes specificationsof desired fruit and seed traits (Anegbeh et al.2003, Leakey et al. 2000, Leakey and Page2006). These traits often exhibit variabilityand, although affected by environmentalconditions into which the plant is exposed,usually reflect the genetic make-up of theindividual, variety or provenance. For farmersplanting jatropha, the size and weight of fruitsmight be the most important expression ofyield as they can easily observe them.

Fruit weight was recorded maximum (16.65g) in SC provenance and minimum (13.58 g) inBC provenance, the mean weight being 15.21g. The fresh weight of fruit is highly affectedby the maturity of the fruit; hence, only fruitsthat had turned yellow were collected for datagathering. If not harvested on time, i.e., thepulp has turned black, the pulp of the fruitsdecays easily and so much loss in weight canbe expected due to the decay of the tissue.

Regarding fruit size, the largest fruits wereobserved in SC provenance while the smallestfruits were from BC. As can be noticed easilythrough visual observation, the fruits werelonger than they were wide and thick. Thefruits were just slightly thicker than they were

wide. As in fruit weight, the SC provenancehad the largest fruits while the BC provenancehad the smallest fruits.

The number of seeds per fruit ranged from2.54 to 2.91. The highest number of seedsper fruit was three although having four seedsis possible. Jongschaap et al. (2007) reportedthat fruits commonly contain three seeds.According to Leakey et al. (2002), a possibleexplanation for variability in seed number perfruit is that not all ovules are successfullypollinated, perhaps because of lack ofpollinators. In the case of the experiment, theplants were regularly spaced.

A web diagram is provided in Figure 2 toshow the relative variability of fruit traits.Researchers who are into ideotypedevelopment are using web diagram tovisualize the relative superiority or inferioritybetween traits and combination of traits(Leakey et al., 2000). From the web diagram,it can be seen that the variability was smallfor every trait. The low variability is alsoindicated by low CVs which ranged from2.24% to 6.44%.

The whole seed weight ranged from 2.95 gto 3.83 g. Lower weight was due mainly tothe condition of having less than three seedsper fruit and not because the seeds differconsiderably in weight (CV = 6.44%).

Seed traits are important variables in treeimprovement programs. In the case ofjatropha, the development of ideotype takesinto account one or a number of seed traits

20

Journal of Society & Technology Vallesteros & Carandang

Figure 2: Web diagram of fruit traits of best trees.

that may lead to increased oil content.Whether seed trait can be targeted forideotype development or breeding depends onthe existence of variability among members ofjatropha populations or among provenances.Variations in seed and seedling traits amongand within seed sources suggest that selectionamong sources might result in rapid geneticgain for the traits (Dangasuk et al., 1997 ascited in Dhanai et al., 2003). However, saidvariability is attributable to either geneticvariability or environmental condition (Leakeyet al., 2002).

Seed traits that were measured in thisstudy include 50-seed weight (fresh andair-dry weights of 50 seeds), 50-kernel weight,whole seed weight, seed length, seed breadth,and seed thickness.

Based on the weight of 50 seeds (65.54 g),the average fresh weight of a seed was 1.31 g.After air-drying to constant weight, theaverage seed weight became 0.76 g, or therewas an apparent loss of moisture amountingto 0.55 g. Another way to express seed weightis by counting all seeds in a fruit. Ginwal etal. (2004) used the term “whole seed weight”

to indicate the fresh weight of all the seeds ofeach fruit, which may number from one tofour. The average whole seed weight that theauthors obtained was 3.49 g. In this study,the highest whole seed weight was observed inTC provenance (3.83 g) and the lowest wasfrom BC (2.95 g).

Oil in jatropha comes from the kernel. Theair-dry weight of 50 kernels ranged from 20 gto 27 g, and averaging 23.92 g. The kernel isaround 62.67% of seed weight.

Measures of size of seeds include seedlength, seed breadth, and seed thickness. Asexpected, seeds were much longer than theywere wide, and thicker than they were wide.Based on a rough calculation (seed size =length x width x thickness), SC provenancehad the largest seeds while BC provenancehad the smallest seeds. The three parametersunder seed size were positively correlated(p<0.01) with one another. Surprisingly, noneof the seed size parameters were significantlycorrelated with 50-seed fresh weight.

The coefficients of variation were low,ranging from 1.5% to 8.6%. Thesetree-to-tree variations are shown in a web

21

Journal of Society & Technology Vallesteros & Carandang

Figure 3: Web diagram of seeds traits of best trees together with oil content.

diagram Figure 3. Seed length had widervariation compared to seed breadth and seedthickness. Still, wider variations wereobserved in seed weight and kernel weight.

According to Kaushik et al. (2007), geneticvariation in seed morphology can be of greatpotential in tree improvement programsparticularly for selection of genotypes havingmore oil content and yield. Kaushik et al.(2007) found significant differences in seedsize (p<0.05) between the 24 accessions of J.curcas from agroclimatic zones of India’sHaryana State. The authors assumed that thevariation in jatropha seed sources on theirseed morphological characters could bebecause the species grows over a wide rangeof rainfall, temperature and soil type. Aslightly larger variation was attributed tophenotypic rather than genetic variables.Dhanai et al. (2003) identified the differencesin nutrients, light, or water in the sites as thepossible cause of variation in seedmorphological traits.

An examination of seed size is not onlyimportant in determining the useful productfrom seeds but also in propagation purposes.Saturnino et al. (2005 as cited in Jongschaap

et al., 2007) found that seed size of jatrophasignificantly influenced the seedling height,rooting depth, stem diameter, number ofleaves and dry matter of green biomass, whichwere all measured 90 days after emergence.Kaushik et al. (2007) reported that bold andheavy seeds of jatropha led to a more vigorousseedlings growth. They suggested that suchperformance would result in lesser nurserymanagement time and maintenance cost.

Correlation betweenmorphological traits

None of the tree form dimensions hadsignificant correlation with any of the fruittraits. The correlation coefficient was eitherpositively or negatively signed, and a trendcould not be drawn out of the relationships.Crown diameter appeared to be a predictor ofthe three seed size parameters (p<0.05),namely: length, breadth and thickness. Thesmaller the crown diameter, the bigger is theseed, suggesting that seeds get enoughnourishment as twigs and leaves are not sothick.

Seed length was significantly correlated

22

Journal of Society & Technology Vallesteros & Carandang

with fruit weight (p<0.01), fruit breadth(p<0.05), and fruit thickness (p<0.01). Also,fruit thickness was positively correlated withseed breadth (p<0.01) and seed thickness(p<0.05). Moreover, the whole seed weightwas positively and significantly correlated(p<0.05) with seed thickness.

Correlation between oil contentand fruit and seed traits

In this study, the product considered forideotype development was the oil content inkernel or the percent crude fat. The averageoil contents in a kernel, based on plot data,were 53.59%, 49.36% and 54.05% for the TL,BC and SC provenances, respectively.Moreover, the overall average oil content was52.33%. The SC provenance had the highestmaximum oil content (59.26%.

Estimates of oil content in kernel vary widelyin literature. The oil content values obtainedin this study were within most of the estimatesfrom other sources.

Correlation between fruits traits and oilyield did not occur, which means that oilcontent could not be predicted using any ofthe fruit traits. Nevertheless, all correlationcoefficients were positive. The highest r value(0.312) was obtained between fruit length andoil content. For ideotype specification,however, large fruits would expectedly be thetrait that farmers would quickly identify for anideotype. They may also specify that fruitsshould always contain three seeds, and laterthey would appreciate such characteristic asuniform ripening of fruits in a bunch.

Seed length was significantly correlated(r=0.498) with oil content. However, seedlength was positively correlated (p<0.01) withseed breadth and seed thickness, which meansthat these traits may be simultaneouslyimproved to increase seed length and later onoil content. Significant positive correlationswere also observed between oil content andair-dry weight of seeds (p<0.05) and betweenoil content and 50-kernel weight (p<0.01). Inthe same manner, Kausik et al. (2007)

observed a positive and significant associationbetween seed weight and oil content. Thesignificant correlation between the saidvariables indicates that the size and weight ofseeds can be improved to enhance the oil yieldof jatropha.

Conclusion andRecommendation

The study provides baseline data on growthperformance and morphological characteristicsof the three provenances. However, there isstill need to consider other provenances withinthe country for superior traits that can beexploited for tree improvement.

Aiming to find traits that can be exploitedin tree improvement program, the studylooked into the relationships of different traitsof J. curcas with its oil content in kernel, theproduct that is sought to be improved. Only afew traits are potentially determinants of oilyield.

Tree improvement may target intermediatetraits such as those of the seeds beforeworking out the most desired products. Thereare prospects with fruit weight, fruit breadthand fruit thickness as shown by theirsignificant correlated with seed length. Also,fruit thickness was positively correlated withseed breadth and seed thickness. Moreover,the whole seed weight was positively andsignificantly correlated with seed thickness.As multi-correlation exists, many traits can besimultaneously improved to enhance yield.

While many ideotype projects target theultimate product that is used through theindustry, directing breeding objective toimprove plant dimensions that farmers cansimply perceive is a good way to go. Farmersmay be more motivated to participate in treebreeding program and may themselves be ableto do simple farm management andsilvicultural activities.

23

Journal of Society & Technology Vallesteros & Carandang

Acknowledgment

The author would like to thank theCommission on Higher Education, and theSoutheast Asian Network for AgroforestryEducation (SEANAFE) which receivedfunding from the Swedish InternationalDevelopment Cooperation Agency (SIDA) forfinancial support.

References

Anegbeh, P. U. C. U. V. T. Z. L. R. .S. K. (2003). Domestication of irvingiagabonensis: 3. phenotypic variation of fruitsand kernels in a nigerian village. AgroforestrySystems, 58:213–218.

Dhanai, CS., U. A. . T. N. (2003). Sourcevariation in albizia chinensis (osbeck) mer.:Seed and seedling . Silvae genetic, 52:259–266.

Foidl, N., F. G. S. M. M. M. . H. S.(1996). Jatropha curcas l. as sourcefor the production of biofuel in nicaragua.Bioresource Technology, 58:77–82.

Ginwal, HS., R. P. . S. R. (2004). Seedsource variation in growth performance andoil yield of jatropha curcas linn. in centralindia. Silvae Genetica, 53:186–192.

Heller, J. (1996). Physic nut. jatropha curcasl. promoting the conservation and use ofunderutilized and neglected crops. Instituteof Plant Genetics and Crop Plant Research,Gatersleben/ International Plant GeneticResources Institute, Rome.

Jongschaap, REE., C. W. B. P. . B. W.(2007). Claims and facts on jatropha

curcas l. Technical report, PlantResearch International, Droevenaalsesteg 1,Wageningen, The Netherlands. Report 158.

Kaushik, N., K. K. K. S. K. N. . R. S. (2007).Genetic variability and divergence studiesin seed traits and oil content of jatropha(jatropha curcas l.) accessions. Biomass andBioenergy, 31:497–502.

Kheira, A. . A. N. (2009). Responsesof jatropha curcas l. to water deficits:Yield, water use efficiency and oilseedcharacteristics. Biomass and, 33 (10):1343–1350.

Leakey, RRB. & Page, T. (2006). The ideotypeconcept and its application to the selectionof cultivars of trees providing agroforestrytree products. Forest, Trees and Livelihood,16:5–16.

Martin, TA., J. K. . W. T. (2001). Ideotypedevelopment in southern pines: Rationaleand strategies for overcoming scale-relatedobstacles. Forest Science, 47 (1):21–28.

RajagopaL, R. (2007). Sustainable biodieselfeedstock: Jatropha: A strategic option. aposition paper by knowgenix in line withjatropha world 2008.

Villancio, V. (2008.). Initiatives for jatrophabiodiesel in the philippines. powerpointpresentation presented during the nationalacademy of science and technology. 30thannual scientific meeting energy securityand sustainability: Assessing the presentand foreseeing the future. Technicalreport, National Academy of Science andTechnology., The Manila Hotel, Manila,Philippines.

24