Chapter 8
Amorphophallus Species - a taxa-specific
approach


8.1 Introduction


The aroid genus Amorphophallus with an estimated number of
species ranging from 90 to 170 primarily occurs in tropical Asia
and Africa (Willis 1955; Bogner et al. 1985; Hay 1988; Hetterschied
pers. comm.). Of these 14 or 15 occur in India, 9 of them, namely,
Amorphophallus bonaccordensis Sivad. & Mohanan, A. bulbifer
(Schott) Blume, A. commutatus (Schott) Engler, A. hohenackeri
(Schott) Engl., A. konkanensis Hett., Yadav & Patil, A. mysorensis
Barnes & Fischer, A. nicolsonianus Sivad., A. paeoniifolius
(Dennst.) Nicolson and A. sylvaticus (Roxb.) Kunth on the hill
chain of the Western Ghats in south India. Of these Amorphophallus
paeoniifolius (Dennst.) Nicolson (syn. A. campanulatus Decne.) is
cultivated over an extensive region, though everywhere on a small
scale for its edible tuber, which is also used in treatment of
digestive disorders. Tubers of a few other Amorphophallus species,
including A. konjac K. Koch, A. muelleri Bl. and A. variabilis Bl.
are considered edible (Bailey 1950). Wild congenerics of the
cultivated Amorphophallus therefore harbor genetic resources of
potential value, especially as the cultivated species are losing
their ability to reproduce sexually because of selective pressures
to preferentially allocate more photosynthate towards tuber
production. The hill chain of the Western Ghats harboring an
estimated 436 species of wild relatives and related taxa of
cultivated plants is of considerable significance for in situ
conservation of these genetic resources. The district of Uttara
Kannada (1352' to 15 30' N lat. and 7405' to 7505' E long.)
with an area of 10200 km2 harbors the most extensive contiguous
tropical humid forest tract of this hill chain with about 191
species of wild relatives of cultivated plants, and is therefore of
particular interest from this perspective. We are involved in an
attempt to work out a strategy for in situ conservation of this
rich diversity of wild relatives of cultivated plants, including
three species in the genus Amorphophallus, namely, A. bulbifer, A.
commutatus and A. paeoniifolius. This chapter reports on the
results of this investigation.
8.2 Data analysis
The data on abundance of Amorphophallus species were analyzed
for density estimation to arrive at preferred habitats. Spatial
distribution patterns were analyzed using the BASIC Programs
POISSON.BAS, NEGBINOM.BAS, BQV.BAS and PQV.BAS provided by the book
"Statistical Ecology" (Ludwig and Reynolds, 1988).
The level of association of Amorphophallus species with other
plant species present in 85 quadrats was computed using Jaccard
index (JI) (Eq. 1) as a measure of similarity between pairs of
species.
a
JIAB = ----------- (Eq. 1)
a + b + c
where,
a = the number of quadrats where both species (A and B) occur.
b = the number of quadrats where species A occurs, but not B.
c = the number of quadrats where species B occurs, but not A.
A complete linkage clustering dendrogram was generated using
these similarity measures with the help of a computer programme
kumar.c developed by us. A rank test was carried out to check
whether the quadrats in which Amorphophallus were present
significantly differed from the total (2300) quadrats with respect
to six parameters as mentioned in chapter 2.
8.3 Results
8.3.1 Abundance, spatial distribution and habitat preference
Table 8.1 shows the abundance of Amorphophallus species in
various habitat types distributed over 46 places. The density of
Amorphophallus in different habitat types is taken as an index of
preference of Amorphophallus toward those habitat types. The
results of the spatial distribution pattern analysis, given in
Table 8.2, make it clear that the spatial distribution of
Amorphophallus species does not follow the poisson distribution,
i.e., they are not randomly distributed. The distribution was
clumped in all the habitats examined. Negative binomial
distribution was found to fit the data quite well as judged by the
chi-square goodness of fit criterion (Table 8.2). The negative
binomial distribution describes clumped, contagious or aggregated
populations. It has two parameters: 1. the mean of the individuals
per sampling unit and 2. a parameter describing degree of clumping.

Table 8.1 Abundance of Amorphophallus species in various habitats.

---------------------------------------------------------------------------
1 2 3 4 5 6 7 8 9
---------------------------------------------------------------------------
Open scrub 1426' 7425' 2 50 22 2200 5067 I
Open scrub 1426' 7424' 1 50 54 10800
Mesa 1426' 7425' 2 80 41 2563
Mesa 1427' 7428' 2 80 0 0
Mesa 1424' 7426' 2 80 54 3375 2125 II
Mesa 1425' 7426' 2 80 41 2563
Acacia auriculiformis pl 1427' 7427' 1 50 0 0
Acacia auriculiformis pl 1425' 7423' 2 50 22 2200
Acacia auriculiformis pl 1426' 7424' 1 50 0 0
Acacia auriculiformis pl 1426' 7423' 2 50 0 0 1767 III
Acacia auriculiformis pl 1421' 7444' 1 50 0 0
Acacia auriculiformis pl 1426' 7424' 1 50 31 6200
Disturbed evg forest 1428' 7427' 1 50 8 1600 1600 IV
Cashew and Eucalyptus pl 1424' 7428' 1 50 4 800 800 V
Casuarina pl 1426' 7423' 2 50 13 1300
Casuarina pl 1426' 7423' 2 50 0 0 650 VI
Areca garden 1429' 7446' 1 50 0 0
Areca garden 1435' 7448' 1 50 0 0
Areca garden 1427' 7429' 1 50 11 2200 550 VII
Areca garden 1418' 7450' 1 50 0 0
Betta land 1419' 7443' 1 50 0 0
Betta land 1421' 7445' 1 50 0 0
Sandy beach and adjoining sand 1422' 7424' 1 50 0 0
Deciduous forest 1425' 7437' 1 50 0 0
Deciduous forest 1426' 7454' 1 50 0 0
Deciduous to semi-evg forest 1423' 7435' 1 50 0 0
Disturbed semi-evg forest 1425' 7430' 1 50 0 0
Disturbed evg to semi-evg forest 1425' 7449' 1 50 0 0
Eucalyptus pl but now polyculture1429' 7433' 1 50 0 0
Eucalyptus pl in evg forest 1430' 7452' 1 50 0 0
Evg to semi-evg forest 1425' 7447' 1 50 0 0
Evg forest 1424' 7444' 1 50 0 0
Evg forest 1416' 7442' 1 50 0 0
Evg forest 1422' 7439' 1 50 0 0
Evg forest 1423' 7439' 1 50 0 0
Evg forest 1420' 7438' 1 50 0 0
Evg forest 1429' 7433' 1 50 0 0
Evg forest 1431' 7434' 1 50 0 0
Evg forest 1423' 7444' 1 50 0 0
Myristica swamp 1416' 7444' 1 50 0 0
Myristica swamp plus evg forest 1425' 7446' 1 50 0 0
Open scrub 1426' 7424' 1 50 0 0
River-side vegetation 1441' 7429' 1 50 0 0
Sandalwood pl 1423' 7455' 1 50 0 0
Teak pl 1426' 7438' 1 50 0 0
Teak pl 1429' 7433' 1 50 0 0
---------------------------------------------------------------------------
Abbreviations: pl = plantation
evg = evergreen
Legends for columns in Table 8.1:
1 = Habitat type
2 = Latitude
3 = Longitude
4 = Quadrat size (in m2)
5 = Number of quadrats
6 = Total No. of individuals
7 = Density per hectare
8 = Over-all density
9 = Order of preference

Table 8.2 Testing goodness of fit for Poisson and Negative Binomial
distributions.
----------------------------------------------------------------------
1 2 3 4 5 6 7 8 9 10 11 12
----------------------------------------------------------------------
Mesa 2 80 41 0.513 6.709 39.33288** 2 NR 0.9297 1 C
Mesa 2 80 41 0.513 3.873 45.66369** 2 NR 5.6499 1 C
Mesa 2 80 54 0.675 6.07 48.20713** 2 NR 2.1657 2 C
Open scrub 2 50 22 0.44 1.027 10.73492** 1 NR 8.6753 2 C
Ac au p 2 50 22 0.44 1.027 8.916841** 1 NR 1.0253 2 C
Ca eq p 2 50 13 0.26 0.727 7.436324** 1 NR 0.1774 1 C
Ac au p 1 50 31 0.62 2.975 21.63098** 2 NR 1.0634 2 C
Minor f 1 50 54 1.08 3.259 49.65658** 3 NR 9.1788 4 C
Dist evg f 1 50 8 0.16 0.994 5.959205* 1 NR 0.4377 1 C
CP & EP 1 50 4 0.08 0.116 5.534192* 1 NR 0.7297 1 C
Areca garden1 50 11 0.22 0.542 3.883375* 1 NR 1.8923 1 C
----------------------------------------------------------------------
* Significant at p=0.05
** Significant at p=0.01

Abbreviations:
f = forest
p = plantation
Ca eq = Casuarina equisetifolia
Ac au = Acacia auriculiformis
CP & EP = Cashew & Eucalyptus plantation
NR = Non random
C = Clumped

Legends for columns in Table 8.2:
1 Habitat type
2 Quadrat size (in m2)
3 Number of quadrats
4 Total number of individuals
5 Mean
6 Variance

For 7 Calculated Chi-square
Poisson 8 Degree of freedom
Distribution 9 Inference

For Negative 10 Calculated Chi-square
Binomial 11 Degree of freedom
Distribution 12 Inference
----------------------------------------------------------------------

Amorphophallus plants were present in 85 of subquadrats
included within the total 2300 quadrats sampled. A rank test showed
that 85 quadrats in which Amorphophallus were present significantly
differed from the total (2300) quadrats in the following
parameters: 1. canopy cover, 2. proportion of three most abundant
plant species <2cm dbh, 3. proportion of three most abundant plant
species >=2cm dbh and 4. proportion of plants >=2cm dbh with an
evergreen phenology. This is also clear from the bar graphs given
in Figure 8.2a, b, c & d respectively. However, the results of the
KS test (Table 4.4) show that quadrats having Amorphophallus
species are significantly different from the total quadrats in the
overall distribution with respect to all the six parameters. Thus
Amorphophallus species prefer habitats with relatively less canopy
cover (10-30%) (Figure 8.2a). Similarly, Amorphophallus species
occur proportionately more in quadrats where plants <2cm and >=2cm
dbh are dominated by few (one to three) species (Figure 8.2b and
8.2c respectively). Also, Amorphophallus plants are proportionately
more in quadrats with less proportion of evergreens, that is, in
disturbed moist deciduous to semi-evergreen portions (Figure 8.2d).
Though the quadrats with Amorphophallus are not significantly
different from total (2300) quadrats with respect to proportion of
exotics in plants >=2cm dbh yet Amorphophallus species are still
relatively more in quadrats with more exotics (Figure 8.2e). This
is because monocultures of exotics are being planted in relatively
open canopied habitats (the preferred habitats of Amorphophallus
which are being replaced). Similarly, the quadrats with
Amorphophallus are not significantly different from total (2300)
quadrats with respect to proportion of tree seedlings in plants
<2cm dbh yet Amorphophallus plants are more in quadrats where
proportion of tree seedlings is 20-30%. Amorphophallus plants are
completely absent from quadrats that are dominated by tree
seedlings (Figure 8.2f).
8.3.2 Species associations
Altogether 64 species (Table 8.4) co-occurred with
Amorphophallus species in the 85 subquadrats. Figure 8.1 represents
a dendrogram depicting the level of association of the more
strongly associated species. Of these two naturally occurring tree
species, Aglaia talbotii Sundararaghavan and Memecylon umbellatum
Burm. were most closely associated, along with a climber Dalbergia
sympathetica Nimmo. These three species are characteristic of open
scrubs. Sapium insigne Benth is another significantly associated
tree species because of its poisonous latex that enables it to
persist in habitats from which most other tree growth has been
eliminated. Several climbers, which thrive in forest clearings are
also significantly associated with Amorphophallus species, as is
the spiny creeper, Mimosa pudica L. characteristic of open
vegetation under heavy pressure of cattle grazing. However, it was
observed that Amorphophallus plants often grow close to some other
plant species like Carissa congesta Wt. (Apocynaceae), Canthium
parviflorum Lam. (Rubiaceae), Strychnos nux-vomica L.
(Loganiaceae), Ervatamia heyneana (Wall.) Cooke (Apocynaceae),
Syzygium species (Myrtaceae) in open habitats like mesa and open
scrubs.

8.3.3 Probable pollinators and the reproductive ecology
We would like to place on record the identity of some insect
species collected from the inflorescence of Amorphophallus
paeoniifolius. A large number of beetles of Adoretus genus were
seen at the base of spadix and spathe. Some bees of Melipona genus
were collecting pollen grains. Two more unidentified insect species
were also getting attracted to the inflorescences of Amorphophallus
species planted in field station. Some insects were occasionally
feeding on the appendix. All insects vanished before withering of
the spathe. Some of such insects might be pollinators of the
Amorphophallus species.
In Amorphophallus bulbifer, A. commutatus (Schott) Engl., and
wild A. paeoniifolius occasionally flowering shoots decayed
without setting seed but in cultivated A. paeoniifolius all the
flowering shoots decayed without setting any fruit. Fresh leafy
shoots emerged from tubers in all (both wild and cultivated) plants
in which flowering shoots had decayed.
8.3.4 Number of seeds per fruit
Table 8.3 shows the frequency distribution of number of seeds
per fruit in Amorphophallus paeoniifolius, A. bulbifer and A.
commutatus. In Amorphophalus paeoniifolius, seeds per fruit ranged
from zero (in failed/undeveloped fruits) to three, rarely four.
The fourth seed, however, was very small. The mode was two but the
mean was 1.98 showing very slightly positively skewed distribution
(Figure 8.3a). In Amorphophallus bulbifer, seeds per fruit ranged
from zero (in failed/undeveloped fruits) to four. The mode was one
but the mean was 1.45 showing highly positively skewed distribution
(Figure 8.3b). In Amorphophallus commutatus, all fruits were single
seeded (Figure 8.3c).

Table 8.3 Frequency distribution of number of seeds per fruit in
Amorphophallus paeoniifolius (Ap), A. bulbifer (Ab), and A.
commutatus (Ac).

+--------------------------------------+
| Frequency distribution |
+-----------+--------------------------+
| No. of | Frequency |
| +--------+--------+--------+
| seeds | Ap | Ab | Ac |
+-----------+--------+--------+--------+
| 0 | 0 | 0 | 0 |
+-----------+--------+--------+--------+
| 1 | 11 | 67 | 73 |
+-----------+--------+--------+--------+
| 2 | 92 | 32 | 0 |
+-----------+--------+--------+--------+
| 3 | 9 | 5 | 0 |
+-----------+--------+--------+--------+
| 4 | 0 | 2 | 0 |
+-----------+--------+--------+--------+
| Total | 112 | 106 | 73 |
+-----------+--------+--------+--------+
| Mean |1.98214 |1.45283 | 1 |
+-----------+--------+--------+--------+
| Variance |0.17825 |0.45532 | 0 |
+-----------+--------+--------+--------+

Note: Failed/undeveloped fruits were not counted.

Table 8.4 List of species which co-occurred with Amorphophallus
species.
-----------------------------------------------------------------------
S. No. Species Family
-----------------------------------------------------------------------
1 Acacia auriculiformis A. Cunn. ex Benth. Fabaceae
2 Acacia catechu (L.f.) Willd. Fabaceae
3 Acanthus species Acanthaceae
4 Aglaia talbotii Sundararaghavan Meliaceae
5 Alseodaphne semecarpifolia Nees Lauraceae
6 Anacardium occidentale L. Anacardiaceae
7 Annona squamosa L. Annonaceae
8 Aporosa lindleyana (Wt.) Baill. Euphorbiaceae
9 Areca catechu L. Arecaceae
10 Bombax ceiba L. Malvaceae
11 Calycopteris floribunda (Roxb.) Poir. Combretaceae
12 Canthium parviflorum Lam. Rubiaceae
13 Carissa carandas L. Apocynaceae
14 Cassia tora L. Fabaceae
15 Casuarina equisetifolia Forst. Casuarinaceae
16 Colocasia species Araceae
17 Costus speciosus (Koenig) Sm. Zingiberaceae
18 Crinum species Amaryllidaceae
19 Curculigo orchioides Gaertn. Amaryllidaceae
20 Curcuma neilgherrensis Wt. Zingiberaceae
21 Curcuma species Zingiberaceae
22 Cyclea peltata (Lam.) Hook. & Thoms. Menispermaceae
23 Cyperus species Cyperaceae
24 Dalbergia sympathetica Nimmo Fabaceae
25 Dioscorea bulbifera L. Dioscoreaceae
26 Dioscorea oppositifolia L. Dioscoreaceae
27 Dioscorea pentaphylla L. Dioscoreaceae
28 Ervatamia heyneana (Wall.) Cooke Apocynaceae
29 Eupatorium odoratum L. Asteraceae
30 Flacourtia sepiaria Roxb. Flacourtiaceae
31 Grewia species Tiliaceae
32 Impatiens species Balsaminaceae
33 Indigofera species Fabaceae
34 Ixora brachiata Roxb. Rubiaceae
35 Ixora coccinea L. Rubiaceae
36 Jasminum species Oleaceae
37 Lannea coromandelica (Houtt.) Merr. Anacardiaceae
38 Laportea interrupta (L.) Chew Urticaceae
39 Memecylon species Melastomataceae
40 Memecylon umbellatum N. Burman Melastomataceae
41 Mimosa pudica L. Mimosaceae
42 Mimusops elengii L. Sapotaceae
43 Musa species (in Areca garden) Musaceae
44 Naregamia alata Wt. & Arn. Meliaceae
45 Nothopodytes foetida (Wt.) Sleumer Olacaceae
46 Ochrocarpus longifolius (Wt.) Benth. & Hook. Guttiferae
47 Pavetta indica L. Rubiaceae
48 Phyllanthus fraternus Webster Euphorbiaceae
49 Piper betle L. (in Areca garden) Piperaceae
50 Sapium insigne Benth. Euphorbiaceae
51 Securinega microcarpa Blume Euphorbiaceae
52 Smilax species Liliaceae
53 Solanum indicum L. Solanaceae
54 Strychnos nux-vomica L. Loganiaceae
55 Syzygium caryophyllatum (L.) Alton Myrtaceae
56 Syzygium cumini (L.) Skeels Myrtaceae
57 Syzygium laetum (Ham.) Gandhi Myrtaceae
58 Syzygium species Myrtaceae
59 Terminalia paniculata Roth Combretaceae
60 Theobroma cacao L. (in Areca garden) Sterculiaceae
61 Theriophonum dalzellii Schott Araceae
62 Vitis indica Wight & Arn. non L. Vitaceae
63 Xeromphis spinosa (Thunb.) Keay Rubiaceae
64 Zizyphus oenoplia Miller Rhamnaceae
-----------------------------------------------------------------------
8.3.5 Number of offset tubers
The number of offset tubers is highly variable across species
as well as within particular species. In Amorphophallus commutatus,
it varied from zero to ten (=4.6, s=2.25, n=68).
The number of foliar bulbils at the major intersections of the
leaflets of Amorphophallus bulbifer ranged from one to 18
depending on the vigor of the plant.
8.3.6 The seed dispersal agents
We would like to record that the two species of bulbuls
(red-whiskered bulbul, Pycnonotus jocosus L. and red-vented bulbul,
Pycnonotus cafer L.) and the koel (Eudynamys scolopacea L.) are
locally well known to favour Amorphophallus fruits, and are snared
at these plants by some people belonging to Mukri and Kari
Vokkaliga communities. Once a very rare case of an Indian myna
(Acridotheres tristis L.) tasting one Amorphophallus fruit was also
seen. A captive male koel consumed hundreds of fruits of
Amorphophallus paeoniifolius, A. bulbifer and Santalum album L.
(Santalaceae) which were offered over 2 days. The koel regurgitated
as well as defecated the seeds. The time gap between last feeding
of the fruits and regurgitation was about 35 minutes and defecation
commenced only later.
8.4 Discussion
The following could be the possible reasons behind the clumped
distribution of Amorphophallus species:
1. Since many fruits of Amorphophallus species are eaten in one
sitting and dispersed by koel and bulbuls, many seeds are
dispersed simultaneously. The upper fruits ripen first, and when
they ripen, they are consumed by these birds. In plants where some
fruits are left behind, they rot and fall near the mother tuber.
Plants growing from these seeds are thus likely to be clumped.
Often, the dispersal agents perch on particular perching sites
repeatedly. Therefore, many seeds fall in the same place. This too
results in a clumped distribution.
2. In Amorphophallus paeoniifolius, the number of seeds per fruit
ranges from zero to three (rarely four) with the mode at two. In
the case of Amorphophallus bulbifer, it ranges from zero to four
while the mode is at one (Table 8.3). Thus, even if a single fruit
is eaten at a time by a bird, the chance that more seeds are
dispersed together is high. However, Amorphophallus commutatus has
only single-seeded fruits, but since its fruits are smaller as
compared to the other two species, they are more likely to be eaten
and dispersed together in greater numbers.
3. Amorphophallus species reproduce asexually as well. Therefore,
more than one leafy shoot can develop from a single tuber. Though
apical dominance is very strong in these plants and usually one
leafy shoot is seen per tuber, some offset tubers often succeed in
producing additional leafy shoots. In Antravalli, it was quite
common to see two leafy shoots coming from a single tuber. There is
much variation in the number of offset tubers per tuber. Sometimes,
the mother tuber rots and the offset tubers start developing
independently. Naturally such plants will be clumped.
Amorphophallus bulbifer produces foliar bulbils at the major
intersections of its leaflets. The number of foliar bulbils varies
from one to 18 per leaf depending on the vigor of the plants. These
foliar bulbils fall fairly close to the mother plant - at a
distance equal to the length of petiole or less. This also
contributes to a clumped distribution.
4. There is polyembryony in some Amorphophallus species like
cultivated A. paeoniifolius (Jos and Vijaya Bai, 1986) and wild
A. paeoniifolius (personal obs.). Therefore, a number of leafy
shoots develop from a single seed in these cases. This again leads
to a clumped distribution.
The association of Amorphophallus species with some of the
other plant species may be related to the fact that its dispersal
agents perch repeatedly on some perching sites provided by other
plant species. Naturally, the Amorphophallus plants coming from
these seeds would seem to be associated with those plants. A number
of climber species are also associated with Amorphophallus species
because they also need some mechanical support. Moreover, these
plant species may also provide some sort of "safe sites" with
better soil, moisture and other ecological conditions and safer
places from being trampled by cattle, human being or other animals.
There are differences in the preferred habitats of different
Amorphophallus species. Amorphophallus commutatus prefers to grow
on mesas where it does well in the scant soil sheltered below
bushes of its associated species and/or between laterite rocks
while A. paeoniifolius prefers habitats where soil conditions are
better. Amorphophallus bulbifer prefers more shade and moist
habitats compared to the other two Amorphophallus species. Apart
from these, there are differences in the preferences of dispersal
agents, mainly birds, for the fruits of different Amorphophallus
species. Amorphophallus commutatus fruits are eaten by both the
dispersal agents (koel and bulbuls) with equal interest since its
fruits are small enough to be swallowed easily by both. But since
the koel does not prefer open habitats like mesas, most of the
seeds of Amorphophallus commutatus on mesas are dispersed by
bulbuls and much less by koels. Both wild and cultivated
Amorphophallus paeoniifolius spadices produce fruits very
systematically arranged on them. Almost all fruits are somewhat
oval-cylindrical except the few lowermost ones. The upper fruits
are smaller and ripen early. Since bulbuls are smaller birds, they
have a smaller gape size as compared to the koel. Therefore,
bulbuls have difficulty in swallowing fruits located at lower
positions of the spadix, but not the koels. When lowermost fruits
are left on very robust spadices of Amorphophallus paeoniifolius
then bulbuls stop visiting them. Then these fruits are dispersed by
koels only. The fruits of Amorphophallus bulbifer are irregularly
shaped and less systematically arranged on the spadix. That is why
they are not much sought out by either of the dispersal agents.
Moreover, the peduncle of this species is more watery, thin, tall,
and weak. When birds perch on it, it tends to bend. These might be
some reasons why Amorphophallus bulbifer is overall dispersed less
effectively and therefore is less common compared to the other two
species. Various Amorphophallus species occur in different patches
with a certain amount of niche overlap and diversification.
As we move from open habitats to the canopied habitats, we
notice a change in the distribution of Amorphophallus species in
that, by and large, both the tuber size and fruit size of the
available species increase with an increase in leaf litter depth
and improvement in soil conditions. Therefore, it is more likely
that Amorphophallus species with big tubers would prefer habitats
where soil conditions are better and leaf litter depth is greater.
For an Amorphophallus seed to germinate and establish itself in an
undisturbed forest, it should have enough food reserves to produce
a long radicle that can penetrate through the deep litter and a
stout plumule to push through the leaf litter above it. This means
that such species should have big seeds (certainly in larger
fruits). To disperse such big seeds, larger dispersal agents with
wider gape and vent opening would be required. Peckover (1985)
reported such dispersal agents from Papua New Guinea, though in
very cautiously-worded sentences because of the limitations of his
observations. Similarly, Sastrapradja et al. (1984) reported that
villagers in different places in West Java had seen bulbuls
(Pycnonotus aurigaster) feeding on the fruits of Amorphophallus
variabilis. It was also reported that the fruits of the
Amorphophallus variabilis were also occasionally fed to the birds
sold in the market in Bogor. Wilbert Hetterscheid (Pers. comm.)
reports that six centimeter large fruits of Amorphophallus titanum
(Becc.) Becc. ex Arcangeli (in Sumatra) are eaten by the large
Hornbills. Obviously, we have an ecological equivalent of this
phenomenon in India as well.
Many fruit-eating birds regurgitate large seeds and pass small
seeds through their alimentary canal unharmed (Howe and Westley,
1988). Since dispersal agents regurgitate as well as defecate, the
defecated seeds would tend to disperse over larger distances
compared to the regurgitated ones because there is a longer time
gap between feeding and defecating than between feeding and
regurgitating. This time gap could be one major factor governing
dispersal distance of seeds from the mother plants. There are
reports of physiological and morphological modifications of the
digestive systems in frugivorous birds and mammals, such as the
shortening of the alimentary canal in both birds and mammals and
the lack of a properly developed gizzard in birds. Hladik (1967)
reports that the obligately fruit-eating spider monkey (Ateles) has
only half the colon surface area as that of a howler monkey
(Alouatta), a leaf-eater of roughly the same size. Similarly,
Walsberg (1975) has reported that a mistletoe specialist bird,
Phainopepla nitens eats the fruits of a mistletoe, Phoradendron
californicum, and can defecate the seeds in as little as 12
minutes. Bulky seeds are regurgitated in even less time. In case of
Santalum album L. (Santalaceae) fruits, which are dispersed by the
same bulbuls and koel as are the Amorphophallus species, the seeds
are defecated or regurgitated depending on their size (Hegde et
al., 1991). Anyway, if there are any morphological or physiological
modifications for speedy processing of fruits of Amorphophallus in
their dispersal agents, this would conceivably tend to reduce the
effective dispersal distances. This could result in a very fast
shrinking of Amorphophallus populations especially if the habitats
of Amorphophallus species and their dispersal agents are allowed to
deteriorate. It must be remembered, however, that the restoration
of Amorphophallus populations would be slow even if such habitats
with its dispersal agents and safe sites for Amorphophallus are
improved. Therefore, for in situ conservation of these plant
species, the habitats of both groups of organisms should not be
allowed to deteriorate to the point of no return.
Populations of Amorphophallus species have shrunken (and
continue to do so) because of large-scale habitat transformation.
The following evidence is a clear indication of declining
populations of Amorphophallus species throughout their range of
distribution:
1. Sastrapradja et al. (1984), while describing edible
Amorphophallus and its related species in Indonesia, wrote the
following:
"With the expansion of agricultural resettlement and other
human activities, the population of Amorphophallus is decreasing,
especially in the densely populated islands such as Java, Madura,
and Bali. Even Amorphophallus campanulatus, which was one of the
most common species, is nowadays considered a strange plant in some
areas because of its unusual inflorescence shape and smell. An
attempt to collect some species from their type localities and
surrounding places has not succeeded yet."
2. Old literature (Hooker, 1897) suggests that Amorphophallus
species (especially A. paeoniifolius) were present right from
Punjab to West Bengal only about 100 years ago. But wild
populations of Amorphophallus paeoniifolius are now either rare
or completely absent in the Gangetic plains.
In our study area (and in other areas of the Western Ghats)
there is human population pressure from adjacent areas - the coast
on the west and the Deccan plateau on the east. Therefore, forests
are shrinking from both sides. Under such a situation
Amorphophallus populations are likely to shrink drastically with
the disappearance of their habitats. At the same time, large-scale
habitat transformations might be adversely affecting the
populations of pollinators and dispersal agents as well. Some of
the wild Amorphophallus inflorescences observed to decay without
setting fruit might be because of lack of pollinators and/or
presence of sterility. But in cultivated Amorphophallus the decay
of all inflorescences might be because of lack of pollinators
coupled with presence of sterility and extreme protogyny. Therefore
long term maintenance of Amorphophallus populations in the wild
would require simultaneous attention to the conservation of
preferred habitats of the plants, as well as the quality of the
habitat for their pollinators and seed dispersal agents.