Provenance Variation

CHAPTER-III

PROVENANCE VARIATION

INTRODUCTION

Provenance, denote the original geographic area from where the seeds or other propagules were obtained. It is one of the major aspect for variation studies and the essence of natural variation involves all of the differences that can be observed or measured between living things (Gupta and Sehgal 1999). For effective utilization of natural variation, the genetic and environmental components are first and foremost to be observed. The provenance or seed source variations in nursery as well as in the field is essentially genetic in origin. The genetic variability depends upon the edaphic and climatic factors operating in the species distributional range. Widely distributed forest tree species show genetic variation between provenances and between individuals of the same provenance. This genetic variation has additive and non-additive components but only the former is exploited in breeding schemes designed to produce superior seeds (Swain et al. 1996). The success of the tree improvement programmes depends upon the determination of amount, type and cause of genetic variability within a species. Proper planning and designing of provenance research enables quick and economical use of those provenances, which yield well adapted and productive forests. Provenance testing is essential to screen the naturally available genetic variation and to choose the best available type for further developing effective tree improvement and breeding work. For any tree improvement programme, it is essential to obtain information about the best provenance for the given site. Plantations will be of low productivity or will fail altogether if provenances and sites are not matched properly (Ladrach 1998). The significance of genetic variation studies and provenance testing in forest tree improvement is very well realized (Meskel and Sinclair 2000).

Environmental factors in combination with genetic and physiological factors play important role in determination of plant potential for seed quality. These characters appear to be under strong genetic control (Roy et al. 2004). Depending on the species, germination responses of seed vary according to geographical factors, viz., latitude, altitude, soil moisture, soil nutrients, temperature, kind and density of plant cover, degree of habitat disturbances of the site where the seed matures. The species having wide geographical distribution may display a wide spectrum of phenotypic variability. So far, only few records of provenance trials of the J. curcas exist where an attempts was made to examine the genetic variation (Heller 1996; Ginwal et al. 2004).

Seeds which are the principal means of propagation contain a lot of variation from one origin to another origin with regards to morphological variation and physiological differences. Knowledge of variability within species is a prerequisite for developing effective tree improvement/breeding strategies (Vakshasya et al. 1992). Seed source variation in pod and seed traits for various forest species is well documented. Seed characters particularly seed weight has been reported to be one of useful criteria for early selection of superior provenances (Khalil 1986) and showed strong correlation with environmental conditions. Gabriel (1978) worked out the relationship between seed traits and geographical factors. Variations in seed size and weight were observed among and within the parent plants (Schmitt et al. 1992).

Variation in seed germination is due to a complex of environmental and genetic factors during seed formation and subsequent handling treatments. Thus, some seed sources showed better performance in certain particular locations than at other locations and the use of seed of a provenance in such areas having very much different climatic and edaphic conditions, often results in poor performance or failure. Variations in germinability of seeds in relation to provenances have been reported in many forest tree species like Albizia lebbek (Kumar and Toky 1993), Pinus pinaster (Falleri 1994), Acacia nilotica (Ginwal et al. 1995), etc.

Ginwal et al. (2005) observed the seed source variation in morphology, germination and seedling growth of J. curcas in central India. Lal et al. (2004) also studied three provenances of J. curcas in U. P., which showed significant variation in seedling traits. Improvement in the species can therefore be achieved through screening of provenances and to choose the best available type for further selection from promising provenances for better seed production and oil content.

METHODOLOGY

The details on selected provenances and different methods for the observations on seed characteristics, germination behaviour and seedling growth for each provenance, adopted in the present study are given below:

Provenance Details

A total of 14 provenances viz., Ahmedabad, Ambikapur, Kottayam, Bhubaneswar, Hyderabad, Indore, Jabalpur, Jhansi, Jodhpur, Coimbatore, Mandsaur, Nagpur, Raipur and Ratlam were selected for the present study. From each of the provenances seeds were collected from ten phenotypicaly strong trees located about 100 m apart from each other in order to avoid narrowing down of the genetic base due to relatedness or inbreeding. Seeds from all the trees of a provenance were mixed and a composite seed lot was made for each provenance.

Seed Characteristics

The observations on morphological characters, moisture content (%) and oil percentage of seeds were made for each provenance.

Morphological Character

To study the variation in seed size traits, eight samples with 25 random undamaged seeds each drawn from each seed lot were measured for their maximum length, width and thickness using an electronic vernier calliper. Fresh seed weight of 8 randomly selected samples, each of 100 seeds for each provenance was determined as per ISTA (1993) rules using electronic top pan balance. The weight of 100 seeds was calculated for each provenance.

Moisture Content (%)

The seeds were oven dried at 800 C ± 20 C to constant weight and the moisture content was determined on fresh weight basis (ISTA 1993):

Fresh seed weight – Dry seed weight

Moisture Content (%) = x 100

Fresh seed weight

Oil Content (%)

Oil was extracted from seeds of J. curcas with the help of Socs plus solvent extractor using acetone as solvent medium in the laboratory.Threeseed samples corresponding to each provenance were oven dried at 80°C for 4-5 hrs, hard seed coat was removed and kernel were crushed into powder. One gram material of crushed kernel was taken for oil extraction. The sample was weighed and placed in the cellulose thimble. Beaker used in this instrument was weighed. Further, 80 ml of solvent (acetone) was added in the beaker. The cellulose thimble containing sample was dipped into the beaker containing solvent. Later, the sample was placed on the instrument and boiled for about 40-45 minutes at 90°C. After 40-45 minutes the temperature of the unit was increased to 150°C and the stopper knob was opened for proper rinsing. Thereafter, the stopper knob closed and solvent was emptied, which was collected in the extractor. Beakers were removed and kept in oven for about 1 hr at 150°C. After 1 hr, these beakers were taken outside, cooled and weighed. The per cent of oil present in a sample was calculated as:

W2 – W1

Oil content (%) = x 100

Where,W1=Initial weight of beaker

W2 = Final weight of beaker (beaker + oil)

W = Weight of powdered sample (1 g)

Germination Behaviour

The germination tests were observed under laboratory as well as in nursery condition also. In laboratory, the experiment was carried out in completely randomized design (CRD) and seeds were kept in the germinator at 270 C temperature by using four replicates of 20 seeds for each provenance. Seeds were placed on sterilized petridishes containing two Whatman germination test papers moistened with distilled water. Observations on seed germination were made every day till the completion of germination. Radicle emergence was taken as an index of germination (Jann and Amen 1977). In nursery, the experiment was laid out in randomized block design (RBD) using 25 bags each in four replications for each provenance. Seeds were sown in polythene bags of 30 x 45 cm size filled with soil, sand and FYM in 3:2:2 ratio. Seeds were considered germinated when sprouted plumules just emerged from the soil surface. The bags were regularly watered and weeded in the nursery. Observations were made daily. The experiment was conducted in the first week of November.

The germination value (GV) was calculated according to Djavanshir and Pourbeik (1976):

GV = (∑ DGS / N x Final cumulative germination % / 10)

where, DGS = daily germination speed

N = number of counts.

Mean germination time (MGT) was computed as per the formula of Yousheng and Sziklai (1985):

MGT = ∑ ni di / n

where, n = total number of germinated seeds,

ni = number of germinated seeds on day di

i = days during germination period.

Speed of germination (SG) was calculated by adopting the Devagiri (1998) method:

A1 A2 A3 An

SG = ------+ ------+ ------+ ------

N1 N2 N3 Nn

where, A1, A2, A3------An = number of seeds newly germinated on

N1, N2, N3------Nnth days.

Germination energy (GE) was calculated, using the Paul (1972) method.

Plant per cent

Plant per cent was observed after 30 days of seed sowing in nursery. It was calculated as:

No. of seedling survived at the end of test

Plant per cent = x 100

Total number of seeds sown

Seedling Growth in Nursery

One hundred seedlings of J. curcas for each of the 14 provenances were raised in the nursery. The experiment was arranged in RBD with four replications. The observations on morphological growth parameters viz., plant height, collar diameter, number of leaves and dry weights of root, stem, branch and leaf were made on ten randomly selected seedlings from each replication at two months interval upto 8 months.

Height of seedlings was measured by graduated tape divided in centimeters and millimeters. Collar diameter of seedlings was measured at collar portion using vernier-calliper and number of leaves per seedling was also counted. At each harvest, seedlings were uprooted after removing the polythene bag and the roots were washed in running water. The root and different components of shoot were separated and kept in hot air oven at 800 C for 2 days. Then final dry weight of different components of seedlings was recorded.

The morphological quality of seedlings of each selected provenance was determined as per Dickson et al. (1960) quality index (QI):

Total seedling dry weight (g)

QI = ------

Height (cm) Top weight (g)

------+ ------

Diameter (mm) Root weight (g)

Seedling Growth in Field

To assess the performance of J. curcas in field condition, 8 months old seedlings of each provenance were transplanted in RBD with three replications at the spacing of 2 x 2 m with 15 seedlings per replication at MES, Kumarganj. The observation on survival per cent was made after one month of plantation. Plant height, collar diameter, number of branches and canopy spread for each provenance were recorded at three months interval for one year, after establishment of seedlings.

DATA ANALYSIS

The data recorded for each parameter were subjected to analysis of variance (ANOVA) as described by Snedecor and Cochran (1967) to test the significance of differences among the provenances. Simple correlations (r) among the traits and between traits and geographical factors of provenance were also calculated following the method of Snedecor and Cochran (1967).

Phenotypic coefficient of variation (PCV) is a measure of total variation present in particular character and was calculated for each seed and seedling related character according to Burton and Devane (1953):

√Vp

PCV (%) = ------x 100

where, Vp = phenotypic variance

X = general mean.

Genotypic coefficient of variation (GCV) which is the measure of total genetic variability present in particular character, was also calculated following Burton and Devane (1953):

√Vg

GCV (%) = ------x 100

where, Vg = genetic variance

X = general mean.

Environmental coefficient of variation (ECV) which is measure of environmental variation existing in a character, was calculated as suggested by Burton (1952):

√Ve

ECV (%) = ------x 100

where, Ve = environmental variance

X = general mean.

Broad sense heritability (h2) is the measure of heritable portion of phenotypic variance and it is the ratio of genotypic variance to the phenotypic variance or total variance. It is a good index of the transmission of character from parents to offsprings. It was estimated by the formula suggested by Burton and Devane (1953) and Johnson et al. (1955):

h2 (%) = ------x 100

where, Vg = genotypic variance

Vp = phenotypic variance.

Genetic gain refers to the average improvement in a progeny over mean of parents and calculated as per Johnson et al. (1955):

Genetic gain = ------x 100

where, GA = genetic advance

X = general mean.

Analysis of Genetic Divergence

Fourteen provenances representing different geographic regions were studied for genetic divergence involving various seed traits collected from different provenances and seedling growth traits recorded both in nursery and field conditions at Kumarganj, viz., seed weight, seed length and width, moisture content, oil content, seedling height, collar diameter, number of leaves and dry weights of different seedling components. The data collected were subjected to non-hierarchical cluster analysis (Beale 1969; Spark 1973) using Euclidean distance on the basis of these quantitative descriptors after standardization.

According to Beale (1969), initially each observation is allocated to its closest cluster centre. The means of the clusters are then calculated and are taken to the new cluster centers. At the same time, the sum of squared deviation of the observations from their respective cluster centre is computed. The observations are then checked in turn to see if a shift to a different cluster centre results in a decrease in the total sum of squares. This assumes that di2 < dk2, where, di is the distance from the centre of cluster i. however, a more effective criterion involves reassigning the observation if the squared deviation from the centre of cluster i is less than that from centres of cluster k, even when the cluster centres are simultaneously repositioned. That is when:

ni nk

------di2 < ------dk2,

ni + 1 nk – 1

where, ‘ni’ is the number of observations in cluster i.

In delimiting clusters usually average deviance among a subset of ‘m’ points is considered, not the individual ½ m (m-1) deviances. If the ith variable on the jth member is Xij , average deviance of a set of ‘m’ is as follows:

1 P m m

------∑ ∑ ∑ (Xij - Xik)2

m (m – 1) i = 1 j = 1 k = 1

1 P m m

= ------∑ ∑ ∑ [(Xij - Xi) – (Xik - Xi)]2

m (m – 1) i = 1 j = 1 k = 1

Where, Xi is the mean of Xi over the ‘m’ members,

1 P

------∑ ∑ ∑ (Xij - Xi)2 + ∑ ∑ - (Xik-Xi)]2 – 2 ∑ ∑

m (m – 1) i = 1 j k j k i k

(Xij - Xi) (Xik-Xi)

The cross-product term vanishes and the other two are equal.

2 m p

Thus, average deviance = ------∑ ∑ (Xij-Xi)2

m – 1 j =1 i =1

Now instead of calculating 1/m (m-1) deviances, ‘m’ deviances from the centre of gravity are calculated.

The assumptions in this method are that the Euclidean distances ‘D’ separating ‘n’ points in a ‘P’ dimensional space are proportional to the dissimilarities between the objects, and secondly, that no object can belong simultaneously to two clusters.

Initially, a given number of vectors of cluster centres are located in the ‘P’ space. The positions of these centres are located in the ‘P’ space. The position of these centres can be chosen arbitrarily or randomly. However, a good choice of initial cluster centres reduces the amount of computation to a considerable extent.

To starts with ‘n’ cases are allotted to predetermined maximum number of clusters (C max.) according to the procedures suggested by Beale. The residual sum of squares, RSS (C) for the solution involving ‘C’ clusters is calculated. Then the number of clusters ‘C’ is reduced by 1 (unless C = min.) and this procedure is repeated till ‘C’ min. is reached, i.e., further reduction is negligibly small. For each step RSS ‘C’ is calculated. When RSS (C) values for C max. ≥ C ≥ C min. are available, these are used in pseudo - F-ratio test of null hypothesis that the solution for the C1 cluster provides no better fit than the solution for the C2 clusters, with C1> C2. This F-ratio is calculated as:

RSS (C2) – RSS (C1) n-C2 C1

F = ------2/P - 1

RSS (C1) n-C1 C2

with P (C2-C1) and P (nn-C1) d.f. The null hypothesis is rejected if the calculated ‘F’ exceeds the table value of F.

For reducing the number of clusters by 1 till C min. is reached, Beale (1969) has suggested certain procedures. Instead of using Beale’s procedure for merging two clusters, Doshi et al. (1981) have adopted a simple procedure. When a solution is found for ‘C’ clusters, ‘C’ vectors of new clusters are calculated. From this set of new cluster centre vectors, last vector is dropped and (C-1) vectors are used as initial vectors of cluster centres for arriving at (C-1) clusters. For determining the appropriate number of clusters, F-test gives a rough guide in exploratory analysis.

RESULTS

The investigations on evaluation of provenances for seed and growth parameters in J. curcas were carried out to study the extent of variation. The results of these studies are described here.

1. Evaluation of Provenances for Seed Characteristics and Germination

(a) Seed Characteristics

Data on seed characteristics (weight, moisture content, length, width, thickness) of 14 provenances of J. curcas are shown in table-3.1. The maximum weight for 100 seeds was recorded for Kottayam provenance (72.58 g) closely followed by Raipur (71.25 g) and Ambikapur (71.05 g) provenances. On the other hand, the seed weight was minimum for Bhubaneswar (39.56 g) and Nagpur (40.32 g) provenances. A highly significant difference was

observed for moisture content among different provenances. However, the differences between Indore and Jabalpur and between Raipur and Ratlam were not significant. Among the provenances, the seed moisture content ranged from 2.12% (Ahmedabad) to 10.57% (Kottayam).

The seed length was recorded maximum for Indore (2.10 mm) and minimum for Jodhpur (1.50 mm). The variation in seed width showed similar trend as for seed length with maximum width (1.30 mm) for Kottayam and minimum (0.80 mm) for Ahmedabad and Ratlam provenances. It was observed that some of the provenances formed separate groups for seed length and width, indicating statistical differences (P< 0.01) among themselves. The thickness of seeds showed low variability among the provenances as compared to that of seed length and width. It ranged from 0.60 mm (Bhubaneswar) to 1.00 mm (Ambikapur and Kottayam). Maximum oil content was recorded for provenances of Chhattisgarh state viz., Raipur and Ambikapur which contained 42.60% and 42.00% oil content, respectively. In the present study 36 % of the total provenances showed seed oil content in the range of 38-39 %, 29 % provenances in the range of 36-37 % and the remaining 21 % provenances showed oil content in the range of 32-34 % (Table-3.1).

(b) Germination Behaviour

Significant variation in germination traits was observed among the provenances of J. curcas (Table-3.2). In laboratory condition, highest germination (91.23 %) was found in the seeds collected from Kottayam and lowest (54.40 %) in Ahmedabad provenance. Eight of the 14 provenances had germination per cent > 80.0 %, three > 60.0 % and in rest of the provenances the germination was comparatively lower. Germination value ranged from 15.37 (Bhubaneswar) to 58.28 (Coimabtore). Maximum speed of seed germination (3.42) was recorded in the seeds collected from Kottayam, while minimum (1.35) was found in the seeds of Mandsaur. The general mean of SG for all the provenances was 2.25. Lowest MGT (7.04) observed in Raipur