NPK Fertilizers for Elephant Foot Yam (Amorphophallus Paeoniifolius (Dennst.) Nicolson) Intercropped with Coffee Trees

Fertilizer application in elephant foot yams (Amorphophallus paeoniifolius (Dennst.) Nicolson) intercropping system is rare in Indonesia, therefore, NPK fertilizers experiment was conducted under the shade of 10-year-old coffee plantation at Leuwikopo Experimental Farm, Bogor, Indonesia, in order to increase the productivity of elephant foot yam intercropped with coffee trees. Prior to planting, 20 ton ha-1 of goat manure was applied. Four NPK combinations, i.e., N, P2O5, K2O at the rate of 0, 0 and 0; 100, 60 and 80; 125, 60 and 100; and 150, 60 and 120 kg ha-1, were applied. Results showed that there were no significant differences in leaf number per plant, petiole size and rachis length among treatments. Application of NPK decreased photosynthetic rates, while increasing rate of N and K2O had no effect on photosynthetic rates. NPK application at the 100 N, 60 P2O5 and 80 K2O kg ha-1 (N100P60K80 treatment) or larger prolonged growth duration regardless of NPK levels, and there was a close relationship between corm yield and growth duration. As a result, corm fresh mass was higher in the 100:60:80 kg ha-1 treatment than in control. In the N125P60K100 and N150P60K120 kg ha-1 treatments, leaves were damaged by heavy rains and winds, counteracting beneficial effect of NPK on growth duration and corm yield. These results suggested the importance of delay of entering dormancy for an increase in productivity of A. paeoniifolius.

In India, elephant foot yam has been commercially cultivated (Srinivas and Ramanathan, 2005). In Indonesia, on the other hand, elephant foot yam is an underutilized crop that commonly grows under shading (Sugiyama and Santosa, 2008). Until 1960s, elephant foot yams had been used as a staple food during the off-season of rice in Central Java. Since the 1960s, the 'Green Revolution' boosted rice production in Indonesia, decreasing the importance of elephant foot yam. However, steamed corms are still used in some districts of Java (Sugiyama and Santosa, 2008).

NPK Fertilizers for Elephant Foot Yam (Amorphophallus
A. paeoniifolius locally called suweg is widely and densely distributed in Java, and is also found in some places of Sumatra, Bali, Lombok and Sulawesi islands (Sugiyama et al., 2010). Ravi et al. (2009) reported that the corm yield of elephant foot yams is 84.6 ton ha -1 and 102.3 ton ha -1 with N application at level of 50 kg ha -1 and 150 kg ha -1 , respectively. However, the productivity of A. paeoniifollius in the field is low in Indonesia, possibly due to bad agricultural practices such as no fertilizer application (Sugiyama and Santosa, 2008).
N and K are important nutrients to increase corm yield of Amorphophallus species (Sugiyama and Santosa, 2008;Ravi et al., 2009;Santosa et al., 2011). Ravi et al. (2009) recommends application of N at levels of 50 to 200 kg ha -1 and K 2 O at level of 75 to 150 kg ha -1 dependent on soil types. Mondal et al. (2012) recommend applying 10 tons manure per hectare for high quality A. paeoniifolius crops. However, there are few studies on the effect of fertilizer on A. paeoniifolius growth under the shade of plantation trees. Singh et al. (2013) stated that the bitter gourd can be profitably grown with A. paeniifolius under multilayer vegetable cropping system, although the corm yield of A. paeniifolius was reduced by about 10% under bitter gourd layer. Sumarwoto (2005) reported that application of P 2 O 5 and K 2 O do not affect growth of A. muelleri under the canopy of Albizia falcataria. Santosa et al. (2006) reported that productivity of A. paeoniifolius growing under 75% shading nets is larger than in 0% shading. Meaning that fertilizers at higher rates is likely required in shading condition than under full sunlight. Hence, we examined the effect of NPK application on the productivity of A. paeoniifolius in coffee plantation.

MATERIALS AND METHODS
The experiment was conducted at Leuwikopo Experimental Farm of IPB, Darmaga (260 m above sea level) from August 2010 to May 2011. Coffee trees (Coffea robusta) were planted in Latosol Darmaga type at a distance 6 m between the rows and 4 m in the rows in 1994 ( Figure  1). By regular pruning, the canopy of coffee tree was maintained at the height of 4-5 m and width of 4.5 m on average. Average light intensity was reduced by 40% of full sunlight. The soil had pH 5.2, and contained a low amount of total N (0.12% by Kjeldahl method), very low amount of Bray I phosphorus (5.9 mg kg -1 ) and low amount of exchangeable potassium (26 me 100 g -1 ). Soil texture was clay (sand:silt:clay = 13.1:22.1:64.8%). Air temperature during experiment was 23 to 32 ºC (25.6 ºC on average) with relative humidity 85-88 %.
Whole one-year-old corms were planted on August 8, 2010. Average seed corm weight was 396 ± 92 g with a diameter of 9 ± 1 cm. Corms were collected from a farmer's field in Kulonprogo, Yogyakarta. According to the farmer's information, the harvested corms were derived from plants without any fertilizer, except for natural tree litter on the farm.
The plots were arranged in completely randomized block design with three replicates. Each plot consisted of one bed (1 m wide and 10 m long) where 15 whole corms with about 1 cm bud were planted in a triangle distance of 60 cm × 60 cm × 60 cm. Land was plowed and harrowed up to a depth of 30 cm using tractor, one month prior to planting. Two weeks prior to planting, goat manure of about 1 kg (equal to 20 ton ha -1 ) was applied in planting hole (25 cm × 25 cm in a depth of 20 cm). Four combinations of N, P 2 O 5 and K 2 O were applied at the rate of 0, 0 and 0 (N 0 P 0 K 0 ); 100, 60 and 80 (N 100 P 60 K 80 ); 125, 60 and 100 (N 125 P 60 K 100 ); and 150, 60 and 120 kg ha -1 (N 150 P 60 K 120 ). N, P 2 O 5 and K 2 O sources were urea (46% N), SP-36 (36% P 2 O 5 ), and KCl (60% K 2 O), respectively.
Prior to planting, corms were immersed to 2% solution of mancozeb. Rice husk was spread out to achieve a thickness of 10 cm at the planting hole. Seed corms were planted with bud face up and planted 10 cm below soil surface. NPK was applied twice; one month after planting at which time bud reached about 15 cm, and two months after planting at which time the first leaf had fully expand. At planting, about 2 g of 3% carbofuran was applied on the cavity of bud. Soil was regularly raised every month. Prior to harvest, daughter corm position was marked by putting stick close to plants.
Watering was applied every three days, if rainfall was less than 20 mm per day each for three consecutive days.
Hand weeding was carried out twice a month. Pesticides, i.e., 2% diazinon and 2% mancozeb were mixed and applied monthly at rate 400 L ha -1 . Plant growth characters, i.e., number of leaves, petiole size, canopy size, photosynthetic rates and time to dormancy, were measured for five plants planted in the center of the beds. Petiole height was measured from 3 cm above soil level to the joint of tripartite rachis. Canopy size was determined by measuring the horizontal distance between the tips of a tripartite leaf and petiole. Photosynthetic rate was measured on February 15, 2011 (at the peak vegetative stage) using LICOR 6400 XT Portable Photosynthetic System. Harvest was carried out after all plants had entered dormancy (May 15, 2011). Dormancy was determined based on complete senescence of leaves.
Daughter corm size and weight were measured. Corms were peeled, sliced, then oven dried at 120 o C for three days for the measurement of dry matter content. Statistical analysis was performed using ANOVA, and mean separation was carried out by Tukey's range test. A chi-square (F ) test was performed to test the fit between a theoretical frequency distribution and a frequency distribution of observed data on corm size and time to dormancy. Analysis of covariance was used to compare the slopes and intercepts of the regression lines.

Leaf Number
First leaf emerged 2 to 4 weeks after planting (WAP), irrespective of NPK treatments. During vegetative growth, plants in all treatments generated more than one leaf from 2 to 6 months after planting (MAP) ( Table 1). At 2 MAP, leaf number was significantly larger in the N 125 P 60 K 100 treatment than in others, but there were no significant differences in leaf number among treatments from 6 MAP and onward (Table 1). This in the contrary with the result of Santosa et al. (2013) that in A. muelleri, application of KNO 3 significantly delayed harvesting time and increased number of leaves.
From 10 WAP to 20 WAP (September to December 2010), monthly rainfall ranged from 178 to 601 mm. On November 10, 2010, there was heavy rain (50 mm day -1 ) and wind up to 43 km h -1 . As a result, leaf number decreased at 4 MAP due to leaf collapse at 14 WAP in the N 125 P 60 K 100 and N 150 P 60 K 100 treatments ( Table 2). Occurrence of leaf collapse increased thereafter in all treatments.

Leaf Size
There were no significant differences in petiole size (length and diamater) and rachis length of both the first and the second leaves among treatments (Table 3). Furthermore, there were no significant differences in canopy size and life span of the first and the second leaves among treatments, although NPK application slightly increased the canopy size and life span of the second leaf (Table 4). In line with this result, Sugiyama and Santosa (2008), and Santosa et al. (2011) have reported that N and K applications increase petiole length.

Photosynthetic Rate
Photosynthetic rates increased with an increase in photosynthetic active radiation (PAR) in any treatment ( Figure 2). PAR ranged from 500 to 1,000 Pmol photon m -2 s -1 under coffee trees during the sunny days. A leaf which first emerges from corm; X Estimated from canopy length and width at the maximum leaf expansion; Y Weeks from emergence to senescence of leaves; Z Mean ± S.E; different letters indicate statistical differences within columns of PAR, photosynthetic rate was 108 to 173 Pmol CO 2 m -2 s -1 with an average of 130 Pmol CO 2 m -2 s -1 . Stomatal conductance ranged from 0.005 to 0.024 mol CO 2 m -2 s -1 with an average of 0.019 mol CO 2 m -2 s -1 .
Photosynthetic rates were significantly higher in control than in NPK treatments at a PAR of 1,000 Pmol photon m -2 s -1 or higher, but there were no significant differences in photosynthetic rates among three NPK treatments, regardless of light intensity (Figure 2). It is considered that actual photosynthetic rates in control plants were almost the same as those in NPK treatments in this study because PAR ranged between 500 and 1,000 Pmol photon m -2 s -1 under the canopy of coffee trees. This might be the reason why there were no differences in leaf size and leaf number among the treatments (Table 1). Gopi et al. (2009) stated that net photosynthetic rate of elephant foot yam increases after triazole (paclobutrazol) application due to increasing in chlorophyll and carotenoid contents.

Daughter Corm
Regarding with the time to dormancy, there were no significant differences among treatments (Table 5), but a large variation was found among plants within the same treatments (Table 6). When chi-square (F ) test was performed using data on the time to dormancy and corm size in the N 100 P 60 K 80 treatment and control, the frequency distribution were significantly different between two treatments; the N 100 P 60 K 80 treatment increased the percentage of corms entering dormancy at 28 WAP or later (Table 6) and also the percentage of corms heavier than 1,000 g as compared to control plants (Table 7).

J. Agron. Indonesia 43 (3) : 257 -263 (2015)
Analysis of covariance indicated that there were no differences in the slopes of the regression lines whether regressions were calculated for each treatment data or for combined data. Figure 3 which was obtained using combined data indicated that fresh mass of corms linearly increased with a delay of harvest. This suggests that yield of A. paeoniifolius was dependent on growth duration, not on the levels of NPK fertilizers (Figure 3). In the N 125 P 60 K 100 and N 150 P 60 K 125 treatments, a high incidence of leaf collapse decreased the percentage of corms which were harvested later than 32 weeks, resulting in low corm weight in the N 125 P 60 K 100 and N 150 P 60 K 125 treatments. Saikia  (2008) reported that early planting is associated with large leaf, thick petiole, heavy corms and prolonged growth duration in A. paeoniifolius, as compared with late planting. This is similar to the present experiment in that prolonged growth period increased yield. Therefore, it is likely that N, P 2 O 5 , and K 2 O applications at the rate of 100 kg, 60 kg and 80 kg ha -1 , respectively, prolongs growth period possibly due to the delay in entering dormancy, leading to larger fresh masses of corms. It is possible that NPK fertilizer could enhance root function in absorbing water for a long time. Santosa et al. (2004) reported that A. paeoniifolius growing in a greenhouse enters dormancy earlier under limited water supply than under sufficient water supply.
The result of the present study showed that growth duration and yield of A. paeoniifolius increased by the application of NPK in addition to 20 ton ha -1 manure, but N, P 2 O 5 , and K 2 O applications at the rate higher than 100 kg, 60 kg and 80 kg ha -1 , respectively, did not show any effect on growth duration and yield. This detrimental finding is likely coincident with collapse leaves due to heavy rainfall and winds. Santosa et al. (2013) stated that application of KNO 3 at the level of 4% favors plant growth and yield when it is applied through foliar than soil application in A. muelleri. Meaning that, collapse leaves resulted in reduction ability of A. paeoniifolius to conduct photosynthesis and to utilize nutrients.

CONCLUSION
NPK application prolonged growth duration, but did not increase photosynthetic rate. Growth duration changed even in the same treatment, causing a large variation in corm size. NPK treatment of 100:60:80 kg ha -1 produced largest corm fresh mass than those of other treatments. Therefore, it seems necessary to elucidate the cultural practices for lengthening growth duration to get higher yield.