Introduction
Adequate fertilization of coconut palm is very important to increase the yield as well as the quality of coconut production in Sri Lanka. Most soils throughout the country are highly weathered and of low fertility. Soil fertility management is thus important and exceptionally so in coconut plantations as coconut is a perennial plant with a continuous productive life. However, mineral fertilizer use in the coconut sector of Sri Lanka is still low and stagnant (Table 1). Although 80-90 per cent of estates (>8 ha) are regularly fertilized, these estates represent only about 18 per cent of the total land area devoted to coconut production. The remaining 82 per cent, is occupied by smallholders (<8 ha) who apply considerably lower fertilizer rates as compared with the average levels indicated in Table 1.

Research and extension activities play a significant role in achieving the national production target of 3,000 million nuts per year, an eight per cent increase over 2006 production levels. The launching of awareness programs and the initiation of field demonstrations among coconut growers at farmers level are the most effective strategies to optimize fertilizer use in coconut.

Materials and methods
In this IPI-CCB project, three demonstration experiments were conducted in 2006-2007 in the following locations (see map below):

1. A wet zone (annual rainfall >2,500 mm): Milathe, Kiridiwala, Gampaha district;
2. An intermediate rainfall zone (1,500-2,500 mm): Olagama, Ridigama, Kurunegala district;
3. A dry zone (<1,500 mm): Mulgirigala, Hambanthota district.

The selected coconut plantations had not been fertilized for a long period of time, thereby allowing the effect of fertilizer application to be revealed clearly after the start of fertilization. Visual nutrient deficiency symptoms, especially N- and K-deficiency symptoms, could be observed in palms in all three locations.

The plot size was about 0.25 ha for each treatment (with ~40 palms per plot). The experimental plots were separated by a guard row to maintain a buffer zone in between treatments. Two fertilizer treatments were applied in each location (Table 2). Fertilizers were applied manually as a single dose application. Other cultural practices were adopted as recommended by the Coconut Research Institute of Sri Lanka. The total nut yield of each treatment was recorded during 2006 and 2007. The average nut yield per palm per year was then calculated.

Soil samples were taken one year after the first fertilizer application (just before the application of fertilizers during the second year).

Results and discussion
Soil physical and chemical properties
Soil texture varied between experimental sites from sandy loam to sandy clay loam (Table 3). Coconut performs well within these soil textural classes if other conditions are good. Soils at the Dry zone site were classified as alfisol whereas those at the other two locations were ultisols. Soils are acidic or slightly acidic, with a low salinity level. Available P (Olsen) is medium to high, available K ranges from low to medium (outside the manure cycle).

Yield response of coconut palm to mineral fertilizers
The majority of palms showing nutrient deficiency symptoms started to recover after three months following the first fertilizer application. Leaves turned from yellow to a more greenish color that was more obvious in treatment 2 (100% recommended dose of APM) as compared to treatment one (50% recommended dose of APM). Palms in the surrounding areas, where fertilizers were not applied, remained under nutrient deficient conditions. A considerable increase in nut yield was observed 3-4 months after the start of fertilizer application in all the locations.

1) Wet zone
The nut yield of coconut obtained during the year 2006 was 43.1 and 54.3 nuts/palm/year in treatments one and two, respectively (Fig. 1). Thus, nut yield increase in treatment two over the treatment one was 26 per cent.

In 2007, nut yield increased from 45.2 nuts/palm/year (treatment one) to 70.5 nuts/palm/year (treatment two). Hence, there was a yield increment of 56 per cent.

2) Intermediate zone
Coconut yield in treatments one and two during 2006 was 37.5 and 52.3 nuts/palm/year, respectively (Fig. 1). This gives a yield increase of 39 per cent in treatment two as compared to treatment one.

In 2007, the yield of nuts in treatments one and two was 51.5 and 73.4 nuts/palm/year, accordingly. Thus, there was 43 per cent nut yield increase.

3) Dry zone
Nut yield in treatments one and two during 2006 was 39.2 and 51.9 nuts/palm/year, respectively (Fig. 1). The yield increase in treatment two over treatment one was 32 per cent. In 2007, nut yield increased from 52.6 nuts/palm/year (treatment1) to 68.8 nuts/palm/year (treatment two). This gives a nut yield increment by 31 per cent.

Conclusions
High fertilizer use efficiency was clearly observed in all the locations as experiments were conducted in the long-term non-fertilized coconut plantations with low soil fertility. Coconut palms strongly responded to the recommended dose of mineral fertilizers compared to the half-recommended rate. The highest average response to mineral fertilizers was observed in the Wet and Intermediate zones of Sri Lanka (41%) followed by the Dry zone (32%). Fertilizer response, in fact, is associated with favorable climate (sunshine hours and rainfall). In the Dry zone, the low rainfall (low soil moisture) is the most critical factor to restricting yield.

The recommended use of mineral fertilizers for coconut is much needed to increase coconut yield and to maintain soil fertility in Sri Lanka.