7. EFFECTS OF PASTURES ON COCONUT YIELDS

This chapter describes in detail the various factors in the pasture-cattle-coconut system influencing coconut yields.

7.1 Effect of species

Aggressive high yielding species have larger nutrient and moisture requirements than less aggressive species of moderate yielding ability. Early work in Sri Lanka (Ferdinandez, 1972; Ramalingam, 1961; Santhirasegaram, 1959, 1964, 1965, 1966a, 1966b, 1975) demonstrated the effect of species on coconut production. Higher forage yields of P. maximum were achieved at the expense of coconuts (see Table 151). No additional fertilizer was applied to pastures. Other work by Santhirasegaram (1964) suggested that B. miliiformis was less aggressive than B. brizantha (see Table 152). According to Guzman and Allo (1975) high yielding grasses such as napier may reduce nut yields up to 40 percent when fertilizer applications are inadequate. Rodrigo (1945) found that without fertilizer napier (P. purpureum) reduced copra yield by 39 percent, but with adequate fertilizer, the yields of copra and fodder were increased by 59 percent and eight-fold respectively.

In Jamaica, pangola grass (D. decumbens) depressed nut yields by 4–9 percent compared with natural pastures (Anon., 1967, 1968), while guinea grass had an even more severe effect (Anon., 1971).

Table 151. - Yields of coconut and herbage DM (Anon., 1982d; Guzman and Allo, 1975; Santhirasegaram 1966c)

Treatment	Number of nuts ha-1 8 years (1958–63)	Herbage DM kg ha-1 8 years (1958–63)
Control (weeds)	8,674	10,895
B. brizantha	9,230	8,714
B. miliiformis	9,203	8,914
P. maximum	8,010	16,447

Table 152. - Effect of pasture species on coconut yield (Santhirasegaram, 1964)

Santhirasegaram (1975) summarized the long term effects of pastures on coconut yields for the period 1956–71. The weed control treatment showed a steady decline in nut yields, while Brachiaria species maintained a yield 15 percent higher than the initial (1956) yield and a total increase of nearly 35 percent compared with the control. P. maximum, a very high N and K demanding species, depressed yields below the control.

Liyanage et al. (1989) demonstrated over a 3 year period that the integration of B. miliiformis/P. phaseoloides/L. Ieucocephala and G. sepium and grazing heifers with coconuts had no ill effects on nut and copra yields of palms in a 45 year old plantation where palms were spaced at 8.4 × 8.4 m (137 palms ha^-1). Jayasundara and Marasinghe (1989) reported that the nut and copra yield of the integrated system was 11 percent higher than on the monoculture system and Liyanage and Dassanayake (1993) noted that by the fourth year there was a 17 and 11 percent increase in nut and copra yield, respectively on the integrated system (see Table 153).

Table 153. - Effect of mixed farming on the yield of coconut (Liyanage and Dassanayake, 1993)

Humphreys (1991) suggests that the reduction in coconut yields by some grass species is not simply related to grass yields as shown in the work of Waidyanatha et al. (1984) where higher yielding grasses gave less reduction in the growth of rubber that lower yielding grasses in Sri Lanka.

In the Philippines, Parawan and Ovalo (1987) noted that annual copra production was 2.10 t ha^-1 where the coconuts were intercropped with native forage species, but increased to 2.36 t ha^-1 under a Setaria sphacelata/Centro cover and decreased to 1.56 t ha^-1 under Brachiaria decumbens/Centro. Where sheep at 10 head ha^-1 grazed B. humidicola, P. notatum and B. decumbens pastures no differences in nut yield were observed, but trials were continuing and differences could show up later (Moog et al., 1993).

Data from Western Samoa (Reynolds, 1981) rates species in terms of effect on coconut production in the order: B. mutica > P. maximum > I. aristatum > B. brizantha > B. miliiformis, with B. miliiformis having least effect as in the earlier Sri Lanka investigations (see Table 154).

Table 154. - Coconut number ha^-1 collected in local and improved pastures over two time periods (Reynolds 1981)

7.2 Effect of fertilizer

There is evidence in a number of countries that most soils are not sufficiently fertile to sustain high yields of both coconut and forage. In Sri Lanka responses to both nitrogen and potassium have been demonstrated (see Tables 155–156).

In India, Pillai (1974) demonstrated coconut yield increases of 23 percent within three years after establishing and manuring various forages under coconuts. Sahasranaman et al. (1983) reported a 28 percent increase in coconut yield with mixed farming largely due to manuring and the better nutrient status of the soil with mixed farming practices.

In the West Indies for satisfactory results it was necessary to fertilize both pastures and coconuts (Anon., 1971).

In Western Samoa the response of coconut yields to the application of fertilizer was shown by Reynolds (1978). Guinea-Centro pastures were established in 1972 with only a small starter dose of fertilizer. By late 1975 coconut yields on a non-fertilized guinea-centro pasture were over 20 percent below those on non-fertilized local pasture (Reynolds et al. 1978). Starting in late 1974 annual fertilizer applications were made of 250 kg ha^-1 of 30 percent potassic superphosphate (7% P, 14% K) split into two equal applications in September/October and March/April. Frond production, measured from October 1975 to October 1976, showed an increase resulting from fertilizer application (see Table 157). (Studies by Chacko Mathew and Ramadasan (1975) demonstrated significant positive correlation between annual nut yield per palm and number of leaves on the crown).

Table 155. - Effect of pastures and nitrogen fertilizer on nut and copra yield (Plucknett, 1979; Santhirasegaram 1965)

Table 156. - Effect of pastures and potassium fertilizer on nut and copra yield (Plucknett, 1979; Santhirasegaram, 1965)

Table 157. - The effect of fertilizer on coconut frond production 1975–1976 (Reynolds et al., 1978)

Treatment	coconut frond production palm-1 yr-1
Local pastures (non-fertilized)	14.0
Guinea-centro (fertilized)	15.2

The general trend of coconut production is shown in Figure 200 and Table 158. By early 1978 the coconut yields on guinea-centro pastures exceeded those on local pastures, and expectation was that by late 1978 yields would be at least 12 percent higher and rising steadily.

Table 158. - Numbers of coconuts ha^-1 collected in local and guinea-centro pastures over four time periods (Reynolds et al., 1978)

The depressive effect of low soil fertility and/or rate of fertilizer application on copra yields is illustrated in Table 159.

Table 159. - Relationship between copra yield (in kg ha^-1) and increasing levels of N fertilizer on control and B. miliiformis pastures (Santhirasegaram, 1975)

While application of fertilizer around each coconut tree is still recommended by some it is clear that fertilizer for the trees and pastures can be broadcast (Habana et al., 1972; Sumbak and Best, 1976) except where nitrogen applications may reduce the effectiveness of legumes in grass-legume pastures.

Where soils are deficient in nitrogen and potassium and farmers are unable to purchase inorganic fertilizers because of their high cost, fresh prunings of G. sepium, L. leucocephala and Calliandra spp. can be incorporated into the soil to help meet the nutrient requirements of coconut palms. In Sri Lanka work at CRI has shown that the incorporation of 30 kg of G. sepium loppings into quarter-circle trenches dug around each palm resulted in a 12% increase in nut yield two years later (Gunasekera, 1989; Liyanage, 1993).

Rajaguru (1991) stresses that trials carried out in Sri Lanka at the CRI indicate that if regular fertilization of both coconut and pasture is maintained then the pasture will have no adverse effects on the coconut. Long-term trials have shown that there is usually an improvement in nut yields where coconuts are intercropped with well managed pastures (Liyanage, 1987).

7.3 Effect of moisture

Coconuts grow best with a well distributed rainfall above 1,250 mm. Moisture stress causes abortion of the young inflorescence growing points and affects the crop yield long after a drought period (Child, 1974).

In Chapter 2 (see Figure 18) it was noted that coconut yields in Malaya (Cooke, 1953) and Trinidad (Murray, 1977; Shepherd, 1926: Smith, 1966) were closely related to previous rainfall. Similar observations have been made in India by Balasubramanian (1956) and Lakshmanachar (1953) and in Sri Lanka Abeywardena (1969) stressed that good rainfall distribution throughout the year is necessary for high nut yields. Coomans (1975) showed that water deficit plays an important role in nut yield fluctuations (under non-limiting conditions of sunshine and temperature) in Ivory Coast, Dahomey, Mozambique and in New Hebrides (now Vanuatu), and proposed an approximate curve for the estimation of copra yields from annual water deficit in mm. (see Figure 201).

Figure 200. - Percentage variation in number of coconuts ha^-1 for local and guinea-centro pastures over time (Reynolds et al. 1978).

Figure 201. - Relationship between copra production and water deficit (and coconut genotype). After Coomans, 1975.

Mean annual rainfall, water deficit and the number of raindays for the four stations used were as follows:

Similar relationships between copra yield and rainfall have been noted in Western Samoa where coconut yields on both local and guinea-centro pastures declined over the period 1974 to 1978 due to a series of dry years (see Figure 202). In an area where mean annual rainfall is 2,929 mm, although the difference in yields between local and guinea-centro pastures was apparently due to nutrition (see section 7.2), there appeared to be a direct relationship between mean monthly rainfall for the preceding six month period and mean monthly coconut yield (r² = 0.81, P = 0.10).

In areas where rainfall was more than 2000 mm with very short dry spells no loss of coconut yields was usually recorded due to the presence of pasture and the resulting competition for soil moisture (Santhirasegaram, 1975).

Figure 202. - Monthly average of coconut yields on local and guinea-centro pastures for period September 1974 to May 1978; Western Samoa (Reynolds et al., 1978).

Where coconut trees are grown in drier areas (with marked dry seasons) yields may be more affected by improved pastures. Where rainfall is between 1500–1200 mm and certainly where it is < 1200 mm with a long dry season, then farmers should be aware of the possible soil moisture-pasture-coconut yield interactions. In these circumstances competition with palms can be avoided or reduced by using a grass such as B. miliiformis which tends to stop growing at the onset of a moisture deficit (Lane, 1981).

Effect of grazing

Work in Sri Lanka demonstrated that ungrazed grass had more of a depressive effect on coconut yields than grazed grasses (see Tables 160–161). Where the grass is heavily grazed there is faster recycling of nutrients through the animal and production of manure. In a situation where grasses are harvested by cut-and-carry system there is also likely to be a depressive effect on copra yields (because of the export from the site of nutrients) unless the removed nutrients are returned in the form of manure or fertilizer. Payne (1985) noted that in the Solomon Islands grazing with cattle ensured that the nut pick-up increased from 75 to 90 percent of the total crop.

Table 160. - The effect of grazing on coconut mean yield on B. brizantha pasture in Sri Lanka, 1966–70 (Ferdinandez, 1971)

Table 161. - The effect of grazing intensity and level of manuring on coconut yields
(Guzman and Allo, 1975; Ferdinandez, 1970)

F = fertilizer;
G = grazing;
O = nil;
N = normal;
H = heavy

In 1993, on an on-going University of Ruhuna/Coconut Research Institute research project (Pathirana et al., 1993) at Tennahena estate, Hakmana in Southern Sri Lanka, coconut yields on grazed paddocks (48–57 nuts palm^-1 yr^-1 and 13.3–16.7 kg copra palm^-1 yr^-1) were reported (see Table 162) to be much higher than those on ungrazed paddocks (41 nuts palm^-1 and 11.1 kg copra palm^-1 yr^-1).

Table 162. - Coconut yields of paddocks ungrazed (UG), grazed without straw (G), grazed with straw (GS), grazed with straw and supplements (GSS) (Pathirana et al., 1993).

* Smoked, dry coconut kernel.
^abc Means bearing different superscripts are different (P < 0.05).

Data from Philippines (Moog and Faylon, 1990) confirmed that the growing of forages and grazing livestock under coconuts can have a significant positive effect on nut yields (see Table 163).

Table 163. - Nut yield in grazed and ungrazed pastures under coconuts (Moog and Faylon, 1990)

In Papua New Guinea in West New Britain estates, copra yield started to rise in the late 1970's when there was a change of management technique (see Table 164). The intergrazing of cattle had a very beneficial effect (Ovasuru, 1988).

In Indonesia, Winaya et al. (1983) reported coconut yield increases of 41 percent on native pasture and sown pasture (Brachiaria decumbens, Siratro and Centro) after grazing (Rika, 1986, 1991).

In Malaysia, Wong et al. (1988) used Kambing/Katjang goats at 6 ha^-1 to graze improved tropical grasses under mature coconuts on Bris soil. The yield of coconuts was increased by grazing (but it was not clear if this was due to the fertilizer applied to the pasture, reduced weed competition or the return of organic manures (faeces and urine) to the soil). Kamaruzaman (1988) noted that the growth of rubber trees had a higher increment under sheep grazing than in ungrazed areas.

Table 164. - Copra production in West New Britain estates following introduction of cattle (t ha^-1) (After Ovasuru, 1988)

Year	Garua Estate	Numondo Estate
78/79	0.84	0.81	0.83
79/80	1.06	0.82	0.94
80/81	1.30	0.96	1.13
81/82	1.55	1.18	1.37
82/83	1.68	1.39	1.54
83/84	1.57	1.47	1.52
84/85	1.71	1.54	1.63
85/86	1.73	1.56	1.65
86/87	1.81	1.51	1.66
87/88	1.60	1.45	1.53
	1.49	1.27	1.38

7.5 Effect of grazing system

The data in Table 165 suggest that the grazing system has little effect on nut yields. Probably as long as the grazing system keeps grass short and competition to a minimum, there is likely to be little difference in terms of nut yields. However, as indicated in section 7.7 the introduction of new pasture species and a new grazing system, without changes in nut collection methods, may lead to a reduction in coconut yields.

Table 165. - Effect of grazing system on coconut yields on B. brizantha pastures, Sri Lanka (Plucknett, 1979; Ramalingam, 1961)

7.6 Effect of stocking rate

In Sri Lanka a stocking increment rate was shown to increase copra yields, due apparently to reduced competition from herbage species and an increased rate of nutrient recycling (Ferdinandez, 1970).

Reports on grazing trials in Bali under 60 year old coconut trees stated that coconuts integrated with pasture produced nuts which were bigger and in larger quantities than those without sown pasture (see Table 166; Nitis and Rika, 1978). However, this possibly resulted from the fertilizer rather than a pasture-cattle effect. From Table 167 it is clear that there was also a stocking rate effect in terms of number of nuts and coconut yield which is attributed to nutrient return in the form of dung and urine. For a later part of the same trial Nitis and Rika (1978) and Rika et al., (1981) demonstrated that coconut number and yield were significantly increased at the two higher of four stocking rates; nut number was 62 percent greater and total weight of nuts 46 percent higher. Nut number and weight from the volunteer pasture area were similar to those from the sown pasture area grazed at the two lighter stocking rates (see Table 168). These figures compare with preliminary reports of an 18 percent increase on the sown pasture area (Nitis et al., 1976). It has been suggested that the coconut number and yield increases at higher stocking rates (Rika et al., 1981), confirms the hypothesis that soil nitrogen availability is positively related to stocking rate (Davidson, 1964).

Table 166. - Effect of pasture and cattle on coconut yield 1974–75 (Nitis and Rika, 1978)

Treatment	Nut no. ha-1	nut yield ha-1 month-1 nut size kg ha-1	No. nuts harvested tree-1 year-1
With sown pasture + cattle	316	599	37.9
Without sown pasture + cattle	253	462	30.4

Table 167. - Effect of stocking rate on coconut yield for the period 7 August 1974 – 5 August 1975 (Nitis and Rika, 1978)

Values in the same column followed by different letters are statistically different at P<0.05.

Moog et al. (1993) reported no adverse effect on coconut yield in Bicol Region, Philippines where fertilized Signal grass was grazed at 1.0 and 2.0 head ha^-1. Similarly, no marked differences in nut yield were noted where signal-centro was grazed at stocking rates of 1.0, 2.0 and 3.0 head ha^-1.

Table 168. - Effect of stocking rate on coconut yield for the period May 1975 to December 1976 (Rika et al., 1981)

Values in the same column followed by different letters differ at P<0.05.

Watson and Whiteman (1981) also found no effect of pasture treatment or stocking rates on the number of nuts or yield of copra per tree in the Solomon Islands in the period 1975 to 1978 (see Table 169). Reports on follow-up trials, using rotational grazing systems in the period 1980–83, also indicated no significant effects of stocking rate or pasture on coconut yields (see Table 170; Smith and Whiteman, 1983).

Table 169. - Effects of pastures and stocking rate on total number of coconuts per year, and total copra yield per tree over 3 years of grazing (Watson and Whiteman, 1981)

* S.E. for comparing stocking rates within pastures.

Table 170. - Coconut production of palms undersown with natural and introduced pastures and grazed at 3 different stocking rates (Smith and Whiteman, 1983)

Of more practical significance was the indication that stocking rate (and resulting pasture height) influences coconut pick-up. The percentage of sprouted nuts was highest at the lowest stocking rate (see Table 171) and it took longer to locate nuts in the less heavily grazed pastures (see Table 172). These factors are of major importance to farmers and plantation managers.

Table 171. - Percentage of sprouted nuts at nut pick-up (28 day intervals) from coconuts undersown with natural and introduced pastures rotationally grazed at 3 stocking rates (Smith and Whiteman, 1983)

Pasture	Mean % nuts over 3 years Stocking rates (weaners ha-1)
	1.5	2.5	3.5
Sown	17	13	11	14
Natural	18	13	12	14

Table 172. - Labour used per coconut collected (nuts man-minute^-1) from a stand of coconuts undersown with natural and introduced pastures rotationally grazed at 3 stocking rates (Smith and Whiteman, 1983)

Year	Pasture	Stocking rates (weaners ha-1)
		1.5	2.5	3.5
1981–82	Sown	2.17d	3.16bc	4.23a	3.41
	Natural	3.04c	3.77b	3.75b	3.29
1982–83	Sown	2.12b	2.91a	2.88a	2.64
	Natural	2.51ab	2.93a	3.03a	2.82

Those stocking rate × pasture means for any one year followed by the same letter are not significantly different (P<0.05).

Chen et al. (1978) demonstrated a stocking rate and grazing response in palm oil production which increased from 103 to 141 kg palm^-1 when stocking rate was increased from 1 beast ha^-1 to 2 beast ha^-1 on Panicum maximum and Stylosanthes guianensis cv. Schofield improved pasture. With natural pasture, without grazing, production was just 49 kg palm^-1 (Rika, 1986).

7.7 Effect of nut collection system and height of grass

The introduction of improved pastures usually means that the sward under the coconut tree is taller than when local pastures were grazed by ‘sweepers’. Also if rotational grazing systems are adopted then paddocks may be left ungrazed for periods of 4 to 6 weeks. Grasses like guinea may grow to 1 m in height over this period thus locating fallen nuts is difficult. If coconut collection is carried out after the cattle have grazed the pasture and moved to a new paddock, then location of fallen nuts is easy, provided grazing pressure has produced a well cropped sward (see Figure 203). Where collection is attempted before cattle graze the pasture, then nut pick-up is considerably reduced and, as indicated in section 7.6, labour efficiency is reduced and number of sprouted nuts may increase. (See Figure 204).

In Western Samoa, where the nut collector's wage was based on collecting a certain quota, it was evident that where he could collect from both local and introduced pastures, he tended to concentrate his collection on local pastures (where nuts could more easily be located) rather than tall guinea-centro pastures (Reynolds et al., 1978). Data in Table 173 indicate that a second collection under management supervision resulted in twice as many nuts being found in the tall grass of the guinea-centro pastures as in local pastures. In effect, an apparent 5 percent reduction in nut yield on guinea-centro pastures resulted in a 5 percent increase. In this case adjustment of management system was recommended so that nut collection followed the grazing cycle thus increasing collection efficiency.

Figure 203. - Collecting coconuts from a well-grazed pasture.

Figure 204. - A sprouting coconut missed by the nut collectors.

Where long bamboo poles are used to dislodge nuts or trained monkeys are sent up the trees, then the problems of locating nuts are reduced.

Table 173. - Influence of nut collection on coconut yields in Western Samoa (Reynolds et al., 1978)

7.8 Effect of legume introduction

The introduction of a legume into a sward established beneath coconut trees may affect the coconut yields. With a pure grass sward the farmer may employ similar management methods to those used with local pastures, but to retain a good grass-legume pasture must both ensure good soil P and K levels (possible need to fertilize regularly) and practice good grazing management. With regular fertilizing the fixed N should benefit coconut yields.

The replacement of a mat forming grass Chrysopogon acicultatus in one of the Lever Brothers estates in the Solomon Islands, by the legume Desmodium triflorum had a marked effect on copra yield (Carrad, 1977). In 1929 the grass had spread to most parts of the Pepesala estate. Cultivation started to replace it with Desmodium triflorum and by 1934 the grass was practically replaced. The yield increase (see Table 174) was certainly significant, three factors may have been responsible: nitrogen-fixation by the legume, a less competitive species or the cultivation effect resulting in increased aeration and nutrient release, or a combination of all three.

Table 174. - The influence of legume (Desmodium triflorum) establishment on copra yield (Carrad, 1977)

7.9 Effect of weed control or up-keep methods

Reported trials demonstrate the effect that up-keep or weed control methods can have on coconut production (Hew Choy Kean, 1978). These data (see Table 175) demonstrate the effect of cultivation, if done incorrectly or too intensively, cultivation can damage the roots of coconut trees. The depressive effect of mechanical up-keep methods appears to depend upon the depth of cultivation, as disc harrowing, to a depth of about 9 in. (23 cm) had the most depressive effect, rotavation to a depth of 4–6 in. (10–15 cm) had a moderate depressive effect while use of the Kedah roller, where the blades sink about 3–4 in. (7.5–10 cm) into the ground, had the least effect. Although sprayed plots outyielded the next best treatment (slashing) by 17 percent, and harrowing and rotavation by about 25 percent, spraying has the disadvantage of exposing bare soil to erosion, exposing coconut roots and can lead to increased insect attacks on the coconut palms in the absence of ground vegetation. Furthermore it is unlikely that this method could be used as a long term control method at smallholder level. Although no costs were presented, comparison between these data and those from the Solomon Islands quoted by Carrad (1977) indicates that grazing cattle probably give yields between hand slashing and spraying (see Table 176). Barnes and Evans (1971) described a trial in Malaysia where use of herbicide increased copra yield by 20 percent over slashing.

Chen (1991) notes that a long term cattle grazing trial under oil palm in Malaysia had a cumulative effect on weeding by improving FFB yield by 18 percent over 3 years when compared with the control where only circle weeding was carried out. Some of the most detailed studies have been carried out under rubber where according to Tajuddin and Chong (1991) the prime objective of rearing sheep is to reduce the cost of weeding. The grazing of sheep reduces the need to use weedicides and labour to control weeds. Wan Mansor and Tan (1982) reported that the cost of weeding in mature rubber was reduced by an average of 21.7 percent or M$ 7.26 ha^-1 yr^-1 over a four year period. Lim et al. (1983) achieved a saving in weeding costs of M$ 15.0 ha^-1 yr^-1 and Abdul Karim (1990) some 30 percent cost saving under 2–4 year old rubber. Tajuddin and Chong (1991) reduced weed control costs by 18–38 percent (or M$ 16–29 ha^-1 yr^-1). As the rubber industry spends an average of M$ 150 million yr^-1 on weed control the size of the problem and potential savings is clear.

Table 175. - Comparison of up-keep method on number of coconuts produced (Hew Choy Kean, 1978)

7.10 Effect of cultivation

Disagreement exists as to the effect of cultivation on coconuts (Plucknett, 1979). In Sri Lanka, Wijewardene (1954) recommended against cultivation in coconuts, favouring weed control by rotary mowing, whereas Silva (1951) was in favour of careful ploughing to break up compacted soils and shallow tillage or subsoiling (15–20 cm deep) every two years. The data in Table 177 suggest that subsoiling was beneficial for Paspalum commersonii pastures and associated coconuts, but had little effect on cori grass (B. miliiformis) pastures and coconuts. Joachim (1929) refers to the adverse effect of disc harrowing five to eight times per year, while Child (1974) suggests that an accepted practice on many Sri Lankan estates was to disc harrow at the end of the monsoon rains. Thampan (1975) quotes data from the Ivory Coast and India (see Table 178) to illustrate the effects of cultivation (and manuring) on coconut production. Balasubramanian et al. (1985) studied the effects and economics of certain herbicides and ploughing practices on weed growth and nut yield. Over a five year period, the herbicides reduced weed growth but did not improve nut yield. However, ploughing the entire area under the coconut palms improved nut yield by as much as 21% compared to the control (see Table 179).

Table 176. - The effect of different methods of undergrowth control on copra pick-up
(Carrad, 1977)

Table 177. - The after-effect of subsoiling on coconut and pasture yields (Plucknett, 1979; Rajaratnam and Santhirasegaram, 1963)

Table 178. - Effect of intercultivation on coconut yield (Thampan, 1975)

Treatments	Nuts yield palm-1 year-1	No. of bunches	No. of nuts bunch-1
No cultivation + no manuring	15.3	4.4	2.9
Regular cultivation (alone)	48.2	9.8	4.8
Regular cultivation + manuring	57.6	10.7	5.5

Table 179. - Effect of various weed control treatments on yield of coconut (Mean number of nuts per palm)

Source: Balasubramanian et al. (1985).
¹ 1. Control - no cultivation except digging basins for fertilizer application.
   2. Ploughing twice a year, alternate alleys in alternate years.
   3. Ploughing twice a year, two metres on either side of the row of palms.
   4. Ploughing twice a year, entire area.
   5. Weed control by two sprays of dalapon at the rate of 2 kg 200 litres water^-1 acre^-1.
   6. Weed control by two sprays of gramoxone at the rate of one litre mixed with fernoxone 0.5 kg in 200 litres water acre^-1

There is probably a slight initial depression of coconut yields following cultivation to establish a pasture under coconuts, because of the damage to the tree roots. However, if cultivation is shallow then this effect is minimal and likely to be more than offset by response to applied fertilizer and improved nutrition. There is also evidence to suggest that on some heavy soils where compaction is a problem, that cultivation can actually increase yields. Guzman (1974) showed that on farms where draft carabaos grazed native grasses between the coconut trees nut yields changed little with time, whereas ploughing not only increased yields, but had a cumulative effect over a number of years (see Table 180).

Table 180. - Effect of cultivation on the annual yield of coconut trees in the Philippines
(Guzman, 1974)

The adverse effect of compaction due to animal traffic can be minimized by regulating stocking rate and can be improved by occasional cultivation and loosening of the soil. At the Culamen Plantations in Malita, Davao del Sur Mindanao, in the Philippines, subsoiling is practised every 6 years to reduce the effect of soil compaction (Khatib, 1981). Cultivation practises have been found to increase coconut yields by about 10–15 nuts per tree per year in the Philippines (Faylon, 1982). However, cultivation can be excessive and the work of Hew Choy Kean (1978), mentioned above, demonstrates that regular deep cultivation to control weeds can depress coconut yields.

7.11 General conclusions

When improved pastures are first established there is likely to be a slight initial depression in coconut yields due to soil/root disturbance and the nutrient demands of the sown pasture. However, if soil moisture and nutrient levels are adequate, if sufficient nutrients are applied in the form of fertilizer to match the expected offtake of pasture where soil fertility is low, if adequate stocking rates are used, then coconut yields should be unaffected or may even increase. The farmer, however, should not expect to sustain high pasture and nut production with only low inputs. For high offtake a high input system will be required. Species generally recommended in this book allow for a moderate level of inputs and outputs. This chapter has demonstrated that there are various factors such as species, grazing system, stocking rate and cultivation which can affect coconut yields. With good management coconut trees on fertilized improved pastures are likely to outyield coconut trees on non-fertilized local pastures by more than 20 percent. The most relevant contrast, is likely to be between these pastures and the large areas where coconut trees are underlain by unproductive weeds and bushes in which as much as 50 percent or more of the coconuts produced may never be collected. If these areas are cleared, improved and grazed, then a significant increase in nut yields is likely to result simply from a higher pickup percentage.