Growth and development of true sago palm (Metroxylon sagu Rottbøll) : with special reference to accumulation of starch in the trunk : a study on morphology, genetic variation and ecophysiology, and their implications for cultivation

Keywords: Metroxylon sagu, Arecaceae, starch crops, plant growth and development, plant morphology, inflorescence structure, electron microscopy, phenological scale, genetic variation, plant taxonomy, folk taxonomy, ethnobotany, leaf area, leaf area index, starch accumulation, starch distribution, plant ecophysiology, tropical lowlands, wetlands, traditional processing, estate cultivation, agronomy, Moluccas, Maluku. True sago palm (Metroxylon sagu Rottbøll) is a stout, clustering palm adapted to swampy tropical lowland conditions. Each axis in a sago palm clump flowers once at the end of its life after having amassed a large amount of starch in its trunk. Man can harvest this starch by felling the trunk, pulverizing the pith and leaching the starch out with water, and use it like other starches for food or non-food purposes. It is a staple food mainly in eastern Indonesia and in Papua New Guinea where it is harvested mostly from semi-managed stands. For establishing sago palm as a full-fledged plantation crop, desirable because of its envisaged large yield potential as a perennial, its niche habitat, and its potential as a raw material provider for bio-ethanol production, the scientific base for establishing the right felling time to harvest the starch needed strengthening. Between October 1988 and November 1990, 27 sago trunks in the Adult Vegetative (AV) or Generative (G) phase belonging to six varieties were selected from semi wild sago stands in the Moluccas, eastern Indonesia: 23 trunks (4 varieties) on the alluvial coastal plain near Hatusua village, Seram Island, and 4 trunks (2 varieties) on hilly terrain near Siri Sori Serani village, Saparua Island. These trunks were felled, dissected, morphologically described and sampled for the amount and distribution of starch they contained. The leafless parts of the trunks were 4.45 to 19.65 m long, had a mean starch density of 4.6 to 254 kg/m3 and contained five to 777 kg of starch (maximum found in a whole trunk: 819 kg). To link starch content to age, the ages of the sampled trunks had to be estimated. To enable age estimation by counting leaf scars on the trunk, the leaf unfolding rate of 36 AV-phase palms around Hatusua (31 palms) and Siri-Sori Serani (5 palms) was monitored for varying periods between 1989 and 1992. Probably due to large variation in habitat and genetic make up, this rate varied from 2 to 14 leaves per year (mean 7.85), rendering number of leaf scars unfit as accurate age estimator. Also trunk height proved unfit for this purpose. From monitoring 5 G-phase palms, the G-phase could be subdivided into 3 sub-phases (G1, G2, G3), recognizable from the ground by the phased development of the successive orders of inflorescence branches. By combining gathered morphological and monitoring data, a phenological scale of a model palm was composed consisting of two parallel timelines of hidden and outwardly visible events: two years after the start of the Establishment (E) phase, the first AV-phase leaf is initiated in the apical growing point, to unfold only 2.5 years later; the initiation of the first AV-phase tissues is followed 12.5 to 14.5 years later by the initiation of the first G-phase tissues, followed 4 to 5.5 years later by the shedding of fruits, and finally by a 2- to 5-year Recycling phase (name proposed here) in which the axis decays and collapses. This scale, which accounts for the large time gap between initiation of trunk parts and their becoming visible, may help to correctly time cultural measures. The 27 sampled trunks could tentatively be ranked according to physiological age into 4 AV phase classes and 9 G phase classes. Since the examined palms belonged to 6 different local varieties, their relative rareness or commonness had to be established to assess the validity of the findings. Based on literature and on interviews with informants, an overview of locally recognised sago palm varieties is presented. The number of unique variety names in 32 localities in Indonesia and Papua New Guinea totalled 325, ranging from 2 (spined vs unspined only) to 34 per locality. On the basis of this survey, the Hatusua varieties were considered average. The nomenclatural category folk variety (fovar, fv.) is proposed to unambiguously name local varieties by adding to the variety name an indication of the location where, and (if known) the ethnic/linguistic group by which that name is used. Leaf area estimation methods were devised to enable investigation of the relationship between leaf area and starch content. In the AV-phase the Total leaf area (TLA) of a sago palm axis ranged from 200 m2 to 325 m2, one axis having an exceptional TLA of 388 m2. The TLA in the G-phase before fruiting mostly remained within the same range, possibly exceeding it for a short period early in that stage. The Leaf area index (LAI) of an individual axis showed an upward trend from 1 - 1.5 in the E-phase to 1.25 - 1.75 in the AV-phase, to more than 2 in the early G-phase, followed by a decrease to about 1.5 again in the late G-phase before fruiting. No fruiting palms were available for analysis. The TLA and LAI of a single trunk could not be linked to the mean starch density of its pith, nor to the total amount of starch the pith contained. Generally, starch density in the trunk first increased with height above ground level, reached a maximum about half-way to two-thirds up the leafless part of the trunk, and then sharply dropped towards the top of the trunk. From the late AV phase onward the maximum starch density ranged from 238 to 284 kg/m3. The four trunks with the highest maximum starch densities, all closely around 280 kg/m3, belonged to three different varieties, suggesting that 280 kg/m3 may be considered the maximum starch storage capacity in the pith of any variety. The starch distribution pattern in the leafless part of the trunk showed a tendency to evolve with age from two tailed (density gradually increasing from base, gradually decreasing towards top) to one tailed (density gradually increasing from base, sharply decreasing towards top). The differences in distribution pattern found strongly suggested that there must be other factors besides age and development phase affecting starch accumulation. Attempts to determine the effect of palm variety and of the environment mostly failed. Potential yield of a model palm based on the maximum starch density of 280 kg/m3 was estimated at 840 kg of dry starch. That this amount is much higher than generally found may partly be due to poor recovery ratios, as the results of a traditionally processed trunk demonstrated: only 47% of the starch in the processed trunk part was recovered, and if the unharvested starch present in the traditionally discarded basal and top part of the trunk is taken into account, recovery drops to 44%. In an attempt to establish the point in time at which a sago palm starts to be a nett consumer of its own starch, the course of the energy producing and consuming capacity of an axis during its life time was modelled based on the assumption that by the end of the AV-phase the existing TLA of the axis produces just the amount of energy needed to maintain existing biomass, to keep up the normal regular growth, and to fill new trunk with starch. Using this model, assimilate requirements for building and maintaining the inflorescence and the fruits could not be met by the production capacity of the leaves plus the starch reserves in the trunk. For this modelling approach to succeed in predicting the turning point from nett production to nett consumption of starch by a sago palm axis, additional data on chemical composition of its parts and on assimilation rate are needed. Lack of precise data on the age of the sampled trunks and lack of uniformity of their genetic make up and growing conditions made it impossible to arrive at the sought-after detailed timetable of the evolution of trunk starch accumulation and depletion to base the right felling time of a sago palm on. The high starch density found in the trunk of a palm with half-grown fruits indicated that depletion of starch reserves by the palm itself may set in much later than generally assumed. Once the course of starch accumulation in time in a single axis is unravelled, the next research question should be how this adds up in a clump - the actual production unit in a plantation - with axes of different age. Timing felling in such a situation should be aimed at maintaining a maximum starch accumulation rate for the plantation as a whole rather than at harvesting a maximum amount of starch per trunk. Data sheets of each palm examined containing all primary and some secondary data, and including photographs, are appended in digital form.