Effect of plant density and initial corm size on the corm yield of konjac (Amorphophallus konjac)

Abstract

The effect on annual konjac corm production of planting heavier corms and increasing plant density was investigated at Pukekohe Research Station, New Zealand. Corm yields were 2.2 times the planted corm weights giving a mean corm weight of 741 g harvested from large corms (325 g) and 334 g from small corms (150 g). The mean corm yield increased 1.5 t ha⁻¹ for every 10 g increase in planted corm weight in the 150–330 g range. Increasing the plant density from 33,330 to 106,670 plants ha⁻¹ had a strong linear effect on corm yield. The total yield of corms and offsets from large planted corms increased by 7.2 t ha⁻¹ for each additional plant m⁻² to lift yields from 34.9 t ha⁻¹ to 88.1 t ha⁻¹. For small planted corms, total yields rose from 13.7 t ha⁻¹ to 43.9 t ha⁻¹ giving an increase of 4.1 t ha⁻¹ for each plant m⁻² increase. Offset production was 12% of the total corm yield grown from large corms and 20% when grown from small corms with no effect from plant spacing.

Keywords:

• Amorphophallus konjac
• corm size
• corm yield
• konjac
• plant density

Introduction

Konjac (Amorphophallus konjac, family Araceae) is a traditional food and medicinal crop in Asia (Kurihara 1979; Liu et al. 1995; Long 1995) that is being investigated as a new economic crop for New Zealand (Douglas et al. 2005). Konjac corms are dried and processed into flour (Tye 1991) for use as a food ingredient or with further refinement to produce glucomannan (Thomas 1997; Takigami 2000). Glucomannan is a high molecular weight polysaccharide with unique rheological and gelling properties that is used in a wide range of food, medical and industrial products (Brown 2000; Zhang et al. 2005). Clinical trials have shown glucomannan lowers cholesterol and blood sugar levels and this has created international interest in using konjac products for the treatment of diabetes and the risk factors associated with coronary heart disease (Vuksan et al. 1999; Chen et al. 2003; Sood et al. 2008). The functional properties of glucomannan and the approval of konjac extracts as food additives in the USA and Europe has increased the demand for konjac products outside Asia (Anon. 2001). This demand has also highlighted the lack of published agronomic information on how to produce konjac corms commercially under Western cropping systems.

In China, konjac is traditionally grown among fruit trees or interplanted between tall row crops that provide shade (Liu et al. 1995; Long 1995). The traditional method of growing konjac in Japan is the Jinenjo method in which konjac is grown on hillsides under shade trees, as a natural, mixed-age population from which large corms are harvested annually (Kurihara 1979). The Uedama method of konjac production is intensive cultivation on flat land in which konjac is grown as an annual crop either to supply corms of a suitable size for subsequent planting or to grow a crop for processing (Kurihara 1979; Follett et al. 2002). Crops are commonly planted in 0.6 m spaced rows with plants spaced 150–350 mm in the row (111,000–47,000 corms ha⁻¹) depending on the age of the corm (Sakai 1983). Planting multiple rows on raised beds or ridges may also be used in Japan depending on grower preference, with corm spacing within rows related to the diameter of the corm with rows 1–1.2 m apart (Follett et al. 2002). Little information has been published on the plant density requirements to maximize the corm yield of konjac. We previously found a large increase in corm yield by lifting the plant density from 28,570 plants ha⁻¹ to 66,670 plants ha⁻¹ but concluded the highest plant density was too low to maximize yields (Douglas et al. 2006). A planting density of 10 corms m⁻² was reported as optimal from konjac trials in France but no trial details were presented (Anon. 2001).

Harvested corms are divided into size classes with smaller corms generally used as planting stock and larger corms marketed (Follett et al. 2002) but no information was found on the size distribution of harvested corms. Planting heavier corms gives higher corm yields (Miura & Watanabe 1985; Douglas et al. 2005) but the increased yields have to be equated against the higher cost of producing the heavier planting stock (Long 1995).

This paper reports the results of a field trial in which we compared the effect of a wide range of plant densities and the use of heavier corms at planting on the corm yield and size distribution of harvested corms from an annual konjac crop.

Materials and methods

The trial was conducted at Pukekohe Research Station on a regularly fertilized (pH 5.9, P 65, K 20, Ca 7, Mg 14) Patumahoe clay loam (allophanic oxidic granular soil) fertilized with 0.7 t/ha of a fertilizer mix containing 10% P, 23% K and 5% S prior to planting. It was designed as a randomized block with four replicates, with the replicates split to compare corm weight at planting as the whole-replicate factor (Table 1) and planting density varying within replicates (Table 2). Ten plant densities were compared planted on 1.5 m wide moulded raised beds with individual plots having either three rows spaced at 0.5 m or four rows spaced at 0.375 m with five intra-row spacings (Table 2). Two treatments had the same plant density (5.33 pl m⁻²), planted in either 0.5 or 0.375 m spaced rows. Corms were hand planted and harvested from 3 × 1.5 m datum plots with a 1 m buffer area at the plot ends planted with the same corm spacing. The corms for planting came from corms stored from the previous season in cool, dry conditions being progeny from an unnamed commercial line imported from Japan in 1994. Healthy undamaged corms were weighed individually and graded into large and small corms with very large corms (>500 g) and very small corms (<100 g) discarded to provide more even lines for planting. Corms were weighed on a row basis and hand planted on 16–18 November 2005 with a soil covering of 50 mm. Weed control was achieved with Sylon® (3 L ha⁻¹) and Roundup® (4 L ha⁻¹) applied pre-crop emergence followed by hand weeding as required until crop senescence. No disease control was required. A side dressing of 30 kg ha⁻¹ N as calcium ammonium nitrate was applied on 18 January 2006. Overhead sprinkler irrigation was used to apply 40 mm water on 14 January, 20 February and 6 March. Rainfall over the trial period was 36 mm (November 2005), 104 mm (December), 61 mm (January 2006), 6 mm (February), 43 mm (March), 142 mm (April). The trial was harvested in August and September 2006 when the corms were dormant with individual corm and offset weights recorded from each row.

Table 1 Konjac corm weight treatments planted in blocks.

Replicate	Average corm weight (g)	Range (g)
1 Large corms	320	230–500
2 Small corms	160	150–230
3 Large corms	330	180–500
4 Small corms	140	100–180

Table 2 Konjac plant spacing treatments in density trial.

Treatments	Inter-row width (m)	No. rows/plot	Plant spacing in row (m)	Plant density (plants m^−2)
1	0.5	3	0.250	8.000
2	0.5	3	0.300	6.667
3	0.5	3	0.375	5.333
4	0.5	3	0.500	4.000
5	0.5	3	0.600	3.333
6	0.375	4	0.250	10.667
7	0.375	4	0.300	8.889
8	0.375	4	0.375	7.111
9	0.375	4	0.500	5.333
10	0.375	4	0.600	4.444

Statistical methods

Analysis was by GenStat (version 15; VSN International 2012) ANOVA with factorial treatments and a ‘split plot’ error structure. Linear and quadratic contrasts were included in the analysis for the density treatments.

Results

Comparison of the corm yields from individual rows showed that there were no yield differences between the inner and outer rows of individual plots as the plant density was increased (P = 0.24).

Harvested corms from planting small or large corms showed similar weight gains of 2.2 times the planted weight resulting in a mean corm weight of 741 g from planting large corms (average 325 g) and 334 g from planting small corms (average 150 g) (Table 3). Mean corm weights decreased at the higher plant densities where large corms were planted but were not influenced by planting density when small corms were planted. The total yield of corms and offsets increased up to the highest plant densities with a large difference in yield from planting large rather than small corms (Table 3). The total yield from planting large corms increased from 34.9 t ha⁻¹ at a plant density of 3.3 pl m⁻² to 88.1 t ha⁻¹ at a plant density of 10.7 pl m⁻² with the total yield from planting small corms increasing from 13.7 t ha⁻¹ to 43.9 t ha⁻¹ over the same plant density range. The total yields contained an average of 12% offsets from planting large corms and 20% offsets from planting small corms with no significant effect from the plant density treatments. The corm yields showed a strong linear response (overall R² = 0.97) to plant density up to the highest density planted with no significant quadratic component in the regression analyses (Fig. 1). More than 90% of the corm yield grown from large corms came from corms weighing more than 500 g at the low to medium plant densities but with a small decline in the percentage at the higher densities. The yield of corms weighing more than 1000 g was not affected by plant density. More than 70% of the total corm yield from small corms weighed more than 300 g but there was a gradual decline in the percentage of the total yield as the plant density increased (Fig. 1). Approximately 20% of the total yield from small corms came from corms >500 g without any influence from the density treatments (Fig. 1). Changing the spatial distribution of the plants at the plant density of 5.33 pl m⁻² gave a significant increase in yield for the small planted corms when planted in wider spaced rows with less intra-row space (treatment 3 vs treatment 9) but had no significant effect on the yield from large corms (Table 3). Overall, individual treatments whether planted in 0.5 or 0.375 m spaced rows with variable intra-row spacings all fitted into a yield pattern dominated by a strong linear response to plant density (Fig. 1).

Table 3 Mean konjac corm weight harvested from planting large (L) and small (S) corms and the corm and total (corm and offset) yields from 10 plant density treatments.

		Mean corm weight (g)		Total weight (t ha^−1)		Corm weight (t ha^−1)
Treatment	Plant density (pl m^−2)	L	S	L	S	L	S
5	3.3	762	294	34.9	13.7	30.9	11.0
4	4.0	746	320	34.5	16.4	29.8	13.1
10	4.4	880	368	46.4	21.1	41.2	17.2
3	5.3	865	437	56.1	29.84	50.1	24.33
9	5.3	744	307	50.2	22.59	44.6	18.07
2	6.7	703	300	54.4	26.8	47.7	21.2
8	7.1	709	323	60.7	32.0	53.4	25.5
1	8.0	654	349	66.8	36.6	59.1	28.3
7	8.9	654	317	70.3	39.5	62.2	31.9
6	10.7	694	323	88.1	43.9	78.6	35.1
Mean		741	334	56.2	28.2	49.8	22.6
SED		73	29	4.9	2.6	4.5	2.1

Figure 1 Total konjac corm yield and corm yields >300 g, >500 g and >1000 g from planting large and small corms at nine plant densities (overall R² = 0.97).

Discussion

Konjac is grown in Asia using traditional cropping systems (Kurihara 1979; Liu et al. 1995; Long 1995). No published information on the density requirements of corm production was found apart from our earlier work which showed that plant densities should be increased above 66,000 pl ha⁻¹ to maximize yields (Douglas et al. 2006). This present study shows that corm yields are markedly increased by lifting the plant density to over 100,000 pl ha⁻¹, with the relationship remaining strongly linear and providing no evidence that the yield increase was declining towards an asymptote. To fully understand the entire corm yield–plant density relationship of konjac there is a need to evaluate still higher plant densities.

The size of corms used at planting had a large effect on the subsequent crop yield, confirming previous results (Miura & Watanabe 1985; Douglas et al. 2005, 2006). To achieve high corm yields requires both high plant densities and heavy initial corms but whether this is the best approach to commercial production remains uncertain. Under the intensive Japanese production methods, offsets and small corms are grown as a primary crop to size them up for the subsequent main crop planting (Kurihara 1979; Follett et al. 2002). The second year crop is harvested and processed for flour (Takigami 2000). This two-stage production system would fit into northern New Zealand annual cropping systems as it is similar to growing potatoes. It is unclear from the literature what the optimum target corm weight should be to produce high quality corms for processing and whether the age of the corm is also relevant. Corms over 300 g are marketed in Japan (Follett et al. 2002) which equate to 2-year-old corms (Takigami 2000). The similar corm weight increase of 2.2 times from small or large planted corms indicates that this multiplier could be used as a guide to the corm size that is needed to be planted to achieve a certain size of harvested corms. Brown (2000) reported that konjac corms were grown for 3 years to produce corms of about 2 kg for processing.

Corm production from annual grown crops needs to be compared with the traditional Jinenjo method of konjac production (Kurihara 1979) in which all size grades of corms and offsets are grown on a perennial basis in random mixtures, with an annual harvest of large corms. Such a production method would be possible in northern New Zealand without fear of ground freezing killing the corms. Random mixtures of corms without rows are also likely to grow at higher plant densities than those used in this study, further increasing the possibility of increased yield.

This trial has shown that konjac can be grown successfully as an annual crop in northern New Zealand, with large gains in crop yield from increasing the plant density and the size of the planting stock. This work has shown that konjac is potentially a new industrial crop for New Zealand as a source of nutraceutical products to treat diabetes and risk factors associated with coronary heart disease.

Acknowledgements

We thank Dr R. A. Littler for his input into the revision of this paper.