Tortcon, ~'ol . _?, ~o. 6, pp . 9 - ?-yg1 . 1983 . ~kltl-0101!83 53 .h7- .

00
Printed in Great Britain . 1983 Per~amon P "ns L :d,

MUSCLE PARALYZING EFFECT OF THE JUICE FROM THE

TRUNK OF THE BANANA TREE

YADHU N. SINGH I * and WILtl.a~i F, DR1'DE\3

'School of Natural Resources, University of the South Pacific, Suva . Fiji, and =Department of
Pharmacology, University of Alberta, Edmonton . Albera, Canada

(Accepted jor publication 8 :Mav 1985)

Y. N. $1~GFI and W. F. Daroen . Muscle paralyzing effect ôf the juicy from the trunk of the
banana tree . Toxicon 23, 973 - 981, 1983 . - The effect of an extract from the trunk of the banana
tree (.+Muses sapierttum) was im~estigated in isolated skeletal muscle pre;arations from the chick,
mouse and frog using twitch tension and intracellular recording techniques . The extract produced,
in the same concentration range and after an initial period of twitch au3mentation, paralysis of
skeletal muscle in both directly and indirectly stimulated preparations . It also had a dose-
dependent stimulant effect on the muscle causing a contracture. The ne::romuscular blockade was
reversed by calcium, but only when added before complete paralysis o: the muscle . On the other
hand, neostigmine usually hastened the blockade and aggravated the contracture. The frequency
of the miniature endplate potential in the mouse phrenic nerve-diaphragm preparation greatly
increased initially, declining to an elevated plateau . Effects on quanta( content of endplate
potentials (e .p .p .s) were studied in the transected mouse phrenic nen~;-hemidiaphragm using
trains of e.p .p .s . In the presence of the extract, only a few e.p .p . trains :ould normally be evoked,
probably due to nerve terminal block. When quanta( content could be measured at low
concentrations of the extract, an increase was usually obtained, 41usd; action potentials in the
frog sartorius muscle were decreased in amplitude until no further potentials could be generated .
The results suggest that the nature of the block produced by the extract resembles chat of a potent
local anaesthetic with an initial atypical labilizing effect on cell calcium rather than a conventional
curariform block.

INTRODUCTION
THE USE of poisonous and toxic extracts from plant and animal sources for hunting
animals, catching fish and in tribal warfare is well known. The skin secretion of the
brightly coloured frogs of the Dendrobatidae family is still used as an arrow or dart
poison by natives of western Colombia (ALBUQUERQUE, 1972 ; `'IYERS aRd DALY, 1933) .
In various parts of China arrow poisons have been used for o~~er 2500 ~~ears . The principal
ingredient of many of these poisons is derived from the tubas of .-tronitum species
(BISSET, 1979) . The plant genera Chondodendron and Strrchnos are the main sources of
the well-known curare arrow poisons (MCINTYRE, 1947),
A lesser known arrow poison is that used by certain Pygmy tribes of central and
southern Africa and is made from the juice of the banana tree trunk . According to
HEY~IER (1974) wooden arrows are given a poisonous tip by' driving thzm into the tree
trunk for about ~24 hr . An animal hit by one of these arrows becomes paralyzed but its
flesh can be cooked and safely eaten. PEREIRA et al. (1963) reported that the juice of the

"Present address and address for correspondence: Department of Pharmacology", Uni~er,tty of Alberta, 9-70
Ntedical Sciences Building, Edmonton, Alberta, Canada T6G 2H7.

973
9'4 YADHU N. SINGH and WILLIAivt F. DRYDEN

leaf stem of the banana tree caused smooth muscle spasm and positive cardiac inotropy,
while the trunk juice was claimed by LEE et al. (1980) to produce a non-depolarizing,
irreversible neuromuscular blockade . In an effort to clarify the mechanism of action, we
have investigated the effects of a crude extract of the trunk juice on muscle contractility
and neuromuscular transmission using twitch tension and electrophysiological techniques .
A preliminary account of these results was presented at the 28th meeting of the
Canadian Federation of Biological Societies, Toronto, June 1985.

MATERIALS AND IvfETHODS

Preparation of crude earact

Preliminary twitch tension studies on trunk juice from 4 varieties of banana trees found in Fiji gave very
similar results. All subsequent work was carried out on the cooking variety or plantain (Afusa sapierrum Linn .,
family Ivlusaceae, called vudi in Fijian) grown on the campus of the University of the South Pacific .
Juice squeezed from pieces of the tree trunk was allowed to stand overnight at room temperature (20-25°C)
for the resin to separate . The aqueous layer was centrifuged at 4000 g for 10 min and the precipitate was
discarded. The supernatant was concentrated at 55-60°C using a rotary evaporator until a viscous liquid was
obtained . This, on freeze-drying, yielded a brown, deliquescent powder which was used in all subsequent
experiments .

Twitch tension recording

Biventer cervicis nerve-muscle preparations (GtrtsaoRC and WARRI~ER, 1960) were obtained from chicks
aged 4- l0 days and mounted with a resting tension of approximately 0.5 g in Krebs-Henseleit (KReas and
HENSELEIT, 1932) solution (NaCI, 6.92; KCI, 0.35; MgSO  0.29; CaCI,, 0.28 ; KH,PO 0.16; NaHCO,, 2.1 ;
dextrose, 2.0 g/I) . The solution was maintained at 37°C and gassed with oxygen containing Sao CO,. The
preparations were stimulated indirectly via the motor nerve in the tendon at a frequency of 0.1 Hz, 0.2 msec
pulse width and supramaximal voltage. To assess reversibility of neuromuscular paralysis, an 80-90°.'o block of
twitch height was established and calcium chloride (to a final Ca concentration of 5 mM) or neostigmine (3 NCI)
was added to the tissue bath .
Isolated phrenic nerve-hemidiaphragm preparations (BÜt .eetrrc, 1946) from freshly killed mice (ICR strain,
20-30 g) were mounted with approximately l g tension in the same conditions as used for chick biventer
cervicis preparations . The muscle was stimulated either indirectly via the phrenic nerve (0.2 msec pulse width) or
directly (2 msec pulse width) at a frequency of O.l Hz and supramaximal voltage. Before direct stimulation of
the muscle, neuromuscular transmission was first abolished by o-bungarotoxin (? Ng/ml) or htmicholinium-3
(HC-3) (2 x 10-° M) . «'hen HC-3 was used, the muscle was first indirectly stimulated in the presence of HC-3 at
a frequency of 5 Hz until neuromuscular transmission failed . Experiments were also performed on
hemidiaphragm preparations in which direct muscle stimulation at a frequency of 0.05 Hz was alternated with
indirect stimulation at a frequency of 0.05 Hz.
Contractions were recorded isometrically on a Grass 79 polygraph using Grass FT03 force-displacement
transducers . Experiments were replicated a minimum of three times in different tissues for confirmation .

Intracellular recording
Experiments were performed on isolated mouse phrenic nerve-hemidiaphragm preparations pinned in a
small-volume perfusion chamber and bathed by freshly oxygenated Krebs-Hen~alait solution maintained at
20-21°C. Standard microelectrode techniques were used . Spontaneous transmitter release was monitored by
the frequency and amplitudes of miniature endplate potentials (m .e .p .p .s). For analysis of evoked transmitter
release the muscle was paralyzed by cutting the fibres about 0.5 cm on either sida of the nerve trunk (BAxsr .aD
and LILLEHEIL, 1%S) . Trains of endplate potentials (e .p .p .s) were elicited by stimulating the nerve at 40 Hz for
1 sec . By using the variance and the mean amplitude (Et.n~tQvtsT and Qt;a,sTet., 19h5) of the last 30 e.p .p .s of a
train of 40, the mean quanta! content of e.p .p .s could be calculated . Values of m .e .p .p and e.p.p . amplitudes
were corrected to standard membrane potentials of -70 mV and -50 mV, respactively (KATZ and THESLEFF,
1957), and, where necessary, e.p .p .s amplitudes were also corrected for non-lines: ;t:mmation (~1.aRrt~, 1955).
Muscle action potentials were recorded at 20-21°C from frog muscles following blockade of mechanical
activity by the method Of ESt:ALONA DF Mown et al. (1982) . Sciatic nerve-sartoriu ; muscle preparations from
frogs (Rangpipiers) were paralyzed by immersing them in Ringer solution containing formamide (2 ~1) and then
20-30 min later returning them to normal Ringer solution (NaCI 6.5 ; (~Cl 0 .14: `aF(CO  0.2 ; NaH,PO,, O .OI ;
CaCI,, 0.12 g/1) . A transient decrease of membrane potential from 35 t 4 mV to 6. _" 3 mV occurred . However,
after about 2-2.5 hr in normal Ringer solution, the membrane potential level r~uil :. recovered to 33 t 3 mV,
after which muscle action potentials were generated by stimulating the attached sciatic nerve.
 Banana Trunk Juice and Muscle Paralysis 97 5

Electrophysiological data are presented as means ± S.E .ïvt . of experiments in at least 3 separate preparations .
Statistical analysis of the data was made using the dann-~~~titney U-test (P < 0.05 being regarded as
significant) .

Drugs used
Drugs used were dibucaine hydrochloride, hemicholinium-3, lidocaine, neostigmine methyl sulphate, procaine
hydrochloride (all Sigma) and formamide (Fisher) . c-Bungarotoxin was fraction S, obtained from whole
Bungarus multicinctus venom (Sigma, lot 72C-1720) by iontxchange chromatography (DRYDEN et Ql ., 1974).

RESULTS

Twitch tension recording

At low concentrations (1-3 mg/ml) the extract caused an almost instantaneous
augmentation of indirectly elicited twitch height, reaching a peak within 2-8 min. The
height of twitches then normally decreased but maintained an elevated plateau for up to at
least 30 min (Fig . 1a). At higher concentrations (3 - S mg/ml) augmentation was still
observed, but twitch height subsequently decreased to establish a degree of blockade (Fig .
lb). At the highest concentrations tested (5 - 8 mg/ml) not only were the twitches
abolished, but also a slowly developing contracture occurred which was indicated by an
upward shift of the baseline tension (Fig . lc). The rates of onset of muscle paralysis and
development of contracture were not consistent in all preparations . Also, the amount of
contracture produced in chick muscles was always greater than those in mouse muscles for
the same extract concentration. In all experiments washout of the preparations with
Krebs - Henseleit solution rapidly restored both the baseline tension and the twitches,
although at high concentrations the size of the contractions after recovery. from the block

A
E 2mg/ml Smln

A r
E 3.5mq/ml W

A r
E 8mg/ml yN

FIG . 1 . EFFECT OF BANANA TREE EXTRACT ON INDIRECTLY' STI~it :LATED tH!tA BIVENTER
PREPARATIONS .
The preparations wen stimulated at a frequency of 0.1 Hz and the extract was ad,ied at E in the
concentrations indicated. In (b) and (c) the preparations were washed at W with Krebs - Henseleit
solution .
9'6 YADHU N . SINGH and WILLIA.tiI F . DRYDEN

~I~ li,r ,!
~ : , r :,
,~ .
~~ I ' ~jlli! ;!~ü~~i, I~ !
,
ui,lüili~ÎiÎIIIiÎII ~iiiliii ~~IiIIiÎlIiliilii~lÎlIIIIIJJvId,i .,,, . . .III .I~iii~iii,illi,. y`,j
A r
E 5mg/ml W
29

5 mln

A
E 8 mg/ml

~~~ ~ ~,~,!,I~~
32
m!n
A
D!b 0.02mM W

rI1 65
d mw

A
DIb2mM

FIG. 2 . A COAIPAR[SON OF THE ACTIONS OF THE EXTRACr AND THE LOCAL ANAESTHETIC DIBUC.aI~E IN
THE MOUSE HEMIDIAPHRAGM PREPARATION .
Panels (a) and (c) show the effects of the extract (S mg/ml) and dibucaine (0.02 m`I), respecti~ ely,
on alternate indirect (smaller twitches) and direct contractions (net frequency 0 . I Hz) and reco~~ery
after washout (at ~ with Krebs-Henseleit solution . In (b) and (d) the extras (8 mg/mll and
dibucaine (2 mgt), respectively, were added at the arrows in the absence of muscle stimulation .
Note the differences in onset time and time to peak and recovery in (b) and (d) .

was slightly less than control (Figs lb, c). The extract also produced a blockade of directly
stimulated preparations in the same concentration range that was effective in indirectly
stimulated preparations . The extents of augmentation of twitches in both directly and
indirectly stimulated preparations were also similar (Fig. 2a) .
To assess reversibility of neuromuscular paralysis, an 80-90070 block of twitch height
was established and CaCI, (to a final Ca concentration of 5 mtii) or neostigmine (3 ph'i)
was added to the tissue bath. Raising the calcium concentration fully re~~ersed the
blockade (Fig. 3a), but only when added before complete paralysis of the muscle . On the
other hand, neostigmine usually hastened the blockade of twitches and further aggravated
the devetopment of the contracture (Fig . 3b) .
To assess the nature of the contracture observed in chick biventer preparations, studies
were also done on mouse phrenic nerve - hemidiaphragm preparations . Muscle
stimulation was, however, discontinued before the extract was added to the preparation .
A quickly developing contracture which was followed by a much slo«~er but spontaneous
relaxation and which eventually restored baseline tension was produced by the extract
(3-8 mg/ml) (Fig. 2b) . In other experiments the effect of the extract on KCI contractures
was examined . A response to KCl was first obtained in the absence of direct stimulation
and then after the extract had abolished responses to direct stimulation . In muscle
paralyzing concentrations the extract (3-6 mg/ml) greatly reduced or completely
abolished contractures induced by KCI (3 - 5 x l0'2 M) (Fig . 4) .
 Banana Trunk Juice and Muscle Paralysis 977

FIG. 3. REVERSIBILITY 8Y CALCIUM AND \EOSTIGMINE OF EXTRACT"[NDUCED BLOCKADE OF INDIRECTLY

EVOKED RESPONSES OF THE CHICK HIVENTER PREPARATION .
In each preparation the extract (4.3 mg/ml) was added at E. In (a) calcium concentration of the
bath was doubled to S mM . In (b) neostigmine (3 PM) was added to the bath at Neo and the
preparation was washed at W with Krebs-Henseleit solution.

5 min

2g
J
A1' A A~
KCI E 5mglml KCI
30mM 30mM

FIG. 4. EFFECT OF THE E%TRACTON KCI CONTRACTURE IN THE~70USE HE~t[DIAPHRAG~I PREPARATION.

A control response to KCI (30 mM) was first obtained before the extract (3 mg/mp was added at
E. After complete blockade of twitches to direct stimulation (0.1 Hz) there was no response to KCI
(30 mM).

These results are reminiscent of the actions of local anaesthetics and thus similarities in
the actions of the banana extract and some typical local anaesthetics were investigated in
the mouse hemidiaphragm preparation . Lidocaine (0.2-0.6 mM), procaine
(0.2-0.6 mM) and dibucaine (10-50 uM) all inhibited both directly and indirectly
induced muscle contractions in the same concentration range without producing any
muscle contracture or an initial twitch augmentation (for typical results see Fig . 2c) .
Recovery of twitches after washout of the preparations with Krebs - Henseleit solution
was considerably slower compared with the extract (Figs. 2a,c). For contracture, doses of
local anaesthetics which were about 10 times as great or higher were required . This
contrasted markedly with the banana extract, which produced contracture in
concentrations only slightly higher than those required for complete muscle paralysis . As
with the extract, local anaesthetic induced contracture was followed b~~ a spontaneous
relaxation of tension . However, the time course of this action was very different, being at
least one order of magnitude greater for local anaesthetics (Fig . 2) . Finally, procaine
(0.3 -0.5 mM) and lidocaine (0.5 mM) also reduced or completely abolished KCl induced
contractures .
The mouse hemidiaphragm preparation was also used to assess « hether the twitch
augmentation observed in both directly and indirectly stimulated preparations was due to
facilitated release of neurotransmitter from nerve endings or to a direct action on the
978 YADHU N . SINGH and WILLIAM F . DRYDEN

m . n d unmIIIIli6iNlll!!!!Illllllllll~aluuaaauiunnw~l~il~l~Ilh ~:
HC-3 E
20NM 3.5mglml

Li jÎI,I~I~~ I~I
IIIw .i~IvIIIIIIIGIIIIIIII IIIIIIIIIIIII~II m,nd
=BGT E
2Nglml 1 .5 mglmI

FIG. S. EFFECT OF THE EXTRACT ON D[RECfLY STIMULATED MOUSE HEMIDIAPHRAGM PREPARATIONS .

In (a) control responses to alternate indirect (smaller twitches) and direct stimulation (net
frequency 0 . l Hz) were first obtained, then hemicholinium-3 (20 lint) was added at HC-3 and the
muscle was indirectly (~) stimulated at a frequency of S Hz until transmission failed . In (b) the
muscle was givrn alternate indirect and direct stimulation (net frequency 0 .1 Hz) . At o-BGT, a-
bungarotoxin (2 pg/ml) was added to the preparation . As indicated, the treatments in the left
panels were continued for a further 15 min, then in the right panels the muscle was stimulated
directly (d) at a frequency of O .l Hz and the extract was added at E in the concentrations shown

muscle . In these preparations neuromuscular transmission had previously been abolished

by pretreatment with HC-3 (which depletes nerve endings of acetylcholine) or a-
bungarotoxin (which irreversibly binds to post-functional chalinoceptors). Typical results
for the effect of the extract on direct muscle stimulation are shown in Fig. 5. Low doses
(1- 2.5 mg/mi) caused twitch augmentation while with higher doses (3 - 5 mg/ml) there
was in addition, twitch blockade and a contracture.

Effects on m. e.p.p.s
The effect of the extract on membrane potential, m .e.p.p. amplitud e and frequency was
assessed in mouse hemidiaphragm preparations . In the concentration range tested
(1-4 mg/ml) there was a decline in membrane potential ranging from 20-25% over a
30 min period . There was also a gradual decline in m.e.p.p. amplitud e (corrected to
constant membrane potential) from the time of application, reaching a constant value
after 5 -10 min. This reduction was significant (PC 0.05) for 3 mg/ml and greater (see
Fig. 6 for typical results) . On the other hand, m .e.p.p. frequenc y rose rapidly, reaching a
peak in 4 -10 min, before subsiding to a plateau which was 50 - 300% higher than
control values . For instance, at a dose of 3 mg/ml, m.e.p.p. frequenc y rose from a
control value of 38 t 7 per sec to a peak of 108 t 12 per sec (2830 of control) in 5 min
and then dropped to about 80 per sec (210% of control) for the next 20 min (Fig . 6). In
one experiment, 6 mg/ml gave a peak value of about 950 per sec (2800% of control) . At
this stage twitching in some muscle fibres became evident and the microelectrode was soon
dislodged.

Effects on e.p.p.s
The observed augmentation of twitch height and increase in m .e .p .p . frequency
induced by the juice extract raised the possibility that the extract may also alter evoked
transmitter release. To test this surmise, e.p.p. quantal contest was measured in
transected phrenic nerve - hemidiaphragm preparations . I n muscle paralyzing
concentratiôns (3 mg/ml and higher) the extract caused a 10-20äo decline in the mean
amplitudes of e.p.p.s (corrected to constant membrane potential) and within 2-5 min
 Banana Trunk Juice and Muscle Paralysis 979

300r

ô
'c
0

:~
200
,.r
c
s

100

S 10 15 20 25
Time (mm)

FIG. Ô. EFFECT OF THE EXTRACT ON M.E.P .P . AMPLITUD E (UPPER GRAPHI AYD FRFQC:ENCY (LOWER
GRAPH) IN !riOUSE HEMIDIAPHRAGM PREPARATIOXS .
Each point represents the mean t S .E .M . of 3 separate determinations . The extract (3 mg/ml) was
applied at time = 0 min .

transmission was abolished after only a few trains of e.p.p .s had been evoked . In
concentrations less than 3 mg/ml abolition of e.p.p. trains took longer and it was possible
to measure quantal content in some preparations . The extract caused a dose-dependent
increase in quantal content, e.g. after 10 min exposure quantal content was 108 ± 7% of
control for 1 .5 mg/ml, ! 18 t 10% of control for 2 .0 mg/ml and 125 ± 11 % of control for
2 .5 mg/ml, the changes being significant (P < 0 .05) except at 1 .5 mg ml .

F~jjects on muscle action potentials

The extract produced a reduction in the amplitude of muscle action potentials evoked
by stimulating the sciatic nerve of formamide-treated frog sartorius nerve-muscle
preparations . The rate of reduction was dependent on the dose applizd, although at all
doses tested (3 - S mg/ml) the action potential was eventually abolished. This action was
accompanied by a decrease in the rates of rise and fall of the action potential. As in the
other intracellular experiments, there was a gradual reduction in membrane potential
from -83 t 3 mV to 65 t 4 mV (n = 3) during the course of the recording.

DISCUSSION

The banana plant is a rich source of several pharmacologicall~~ active substances .

Noradrenaline, dopamine, serotonin and an anti-ulcerogenic factor have all been
associated with the skin and pulp of the fruit (WAALKES et al., 1958 ; \'EST, 1958 ; BEST et
al., 1984) . PEREIRA et al. (1963) reported activity in the juice of the leaf items that caused
smooth muscle spasm and positive cardiac inotropy in experimental preparations . The
980 Y:~DHU N . SItiGH and WILLIAM F . DRYDEN

activity of the juice used in the present experiments revealed a series of events more
complicated than the non-depolarizing block envisaged by LEE et al. (1980) .
Since serotonin is a known constituent of bananas, and since TEERAPONG and HAtt~~EY
(1977, 1979) demonstrated that serotonin at millimolar concentrations exerted both
stimulant and inhibitory effects on indirectly stimulated skeletal muscle preparations, it
was necessary to establish if the activity in banana trunk juice was due to its serotonin
content . Three pieces of evidence suggested that this was not so . Firstly, we were unable to
detect significant absorption of u .v, light at 240 and 275 nm which are diagnostic
wavelengths for serotonin . In addition, TEERAPONG and HARVEY (1977, 1979) noted that
no augmentation of twitches occurred with serotonin on directly stimulated muscle, and
up to concentrations of S mM there was no effect on m .e .p .p . frequency . Since we
observed both an initial augmentation of both directly and indirectly elicited twitches and
a profound increase in m .e .p .p . frequency, we conclude that the activity of banana trunk
juice is not attributable to serotonin .
There was no evidence to support the view that the active substances) in the extract
exerted a curarifom block, as suggested by LEE et al. (1980) . Since both directly and
indirectly elicited twitches were equally affected, the site of blockade cannot be the
neuromuscular synapse . An alternative mechanism must be sought . Two alternatives were
therefore considered : either a generalized local anaesthetic effect in which the production
and propagation of action potentials is suppressed in both nerve and muscle, or an
interference with the process of excitation-contraction coupling . The rapidity of onset
and reversal observed with the extract was considered to be uncharacteristic of drugs that
act within the cell and are required to penetrate cell membranes . On the other hand, the
experiments on the suppression of muscle action potentials support the view that the
principal mechanism of action is that of a local anaesthetic . Local anaesthetics suppress
both the directly and indirectly elicited twitches of nerve-muscle preparations with a
similar time course . In addition, high concentrations of local anaesthetics induce a
contracture of skeletal muscle . Local anaesthetics, by also depressing the conductance of
membrane potassium channels, suppress KCl induced contractures (LÜ'I'rGAU and
OETLIKER, 1968) . Since this is also seen with banana juice extract, the argument for a
local anaesthetic mechanism of action is further strengthened .
However, local anaesthetics do not induce an augmentation of either direct or indirect
twitches . The augmentation of twitches which precedes blockade by banana extract is
unlikely to be an anticholinesterase effect, since it occurs in the absence of transmitter
release following hemicholinium exposure and also when postsynaptic receptors have
been irreversibly blocked with a-bungarotoxin . The effect may be related to the
potentiated release of transmitter noted in studies of m .e .p .p . frequenc y and e .p .p .
quantal content . Multiple firing of muscle fibres was not seen in the e .p .p . studies, and so
the enhanced twitch tension is unlikely to be due to summation of individual twitches .
However, both transmitter release and contraction require calcium influx and the
observed effects could be explained by either an increased calcium inflow, or the release of
bound calcium within the cells . Since the time course of the muscle action potential is in
fact prolonged in the presence of the extract, but only in keeping with profound
depression of the action potential amplitude, an enhanced calcium inflow is unlikely to
occur due to prolonged depolarization . This then raises the possibility that the effects are
attributable to labilization of calcium from binding or storage sites within the cell or by
increased calcium permeation of the t-tubule membrane by other means . Thus, increasing
the extracellular calcium concentration will increase the calcium concentration gradient
 Banana Trunk Juice and Muscle Paralysis 9ßl

and facilitate excitation - contraction coupling. The augmented twitch which results will
antagonize the depressant local anaesthetic effect of the extract . It is not known if these
properties reside in one or more substances in the extract and experiments are in progress
to isolate and identify the active substance(s) .
Acknowledgements - This work was supported by the bluscular Dystrophy association of Canada. A travel
grant from the Commonwealth Foundation, London, U.K ., (to Y.N .S .) is gratefully acknowledged .

REFERENCES
ALHI:QUERQUE, E. X . (1972) The mode of action of batrachotoxin . Fedn . Proc . Fedn . Am . Socs. exp . Biol. 31,
1133 .
BARSTAD, J. A. B. and LILLEHEtL, G . (1968) Transversely cut diaphragm preparation from the rat. Archs int.
Pharmacodyn . Ther, 197, 373 .
BEST, R., Lewis, D. A. and NASSeA, N . (1984) The antiulcerogenic activity of the unripe plantain banana (.Muss
species) . Br. J. Pharmac . 82, 107 .
BISSEr, N. G . (1979) Arrow poisons in China. Part 1 . J. Ethnopharmac. 1, 323 .
BI:LBRING, E. (1946) Observations on the isolated phrenic nerve diaphragm preparation of the rat. Br. J.
Pharmac. 1, 3ß .
DRYDEN, W. F., HARVEY, A. L. and MARSHALL, 1. G . (1974) Pharmacological studies on the bungarotoxins :
separation of the fractions and their neuromuscular activity . Eur, J. Pharmac. 26, 236 .
ESCALONA DE MorrA, G., CORDOBA, F., DE LEON, M. and DEL CASTILLO, J . (1982) Inhibitory action of high
formamide concentrations on excitation-contraction coupling in skelaal muscle . J. Nturosci. Res . 7, 163 .
EL~tQvisr, D. and QuASrEL, D. M. J. (1%S) A quantitative study of endplate potentials in isolated human
muscle . J. Physiol., Lond, 178, SOS .
GINSBORG, B. L. and WARRINER, J . (1960) The isolated chick biventer cervicis nerve-muscle preparations . Br.
J. Pharmac. 15, 410 .
HEYAIER, A, (1974) The hunting and food-gathering pygmies. Image Roche. 62, l7 .
ICnrz, B. and THFSLEEF, S . (1957) On the factors which determine the amplitude of the miniature endplate
potential . J. Physlol., Lond. 137, 267.
KREBS, H . A. and HEtvsELerr, K. (1932) Utersuchungen ûber die Hamstoffbildung Tierkorper. Hoppt-5eyler's
Z . physiol. Chem . 210, 33 .
LEE, S. K., NG, L. L. and LEE, S . I . (1980) Experiments with banana trunk juice as a neuromuscular blacker.
Rust . J. exp. Biol. med. Sci. 58, 391 .
LOrrGAU, H. C. and OETLIKER, H. (1%8) The action of caffeine on the activation of the contractile mechanism
in striated muscle fibers . J. Physiol., Lond. 194, S l .
MAATiN, A. R. (1953) A further study of the statistical composition of the endplate potential . !. Physiol., Lond.
130, 114 .
Mc1vrYAe, A. R. (1947) Curare. Its History, Nature and Clinirnl Use. Chicago: University of Chicago Press.
MYERS, C. W . and DALY, J. W. (1983) Dart-poison frogs. Scient. Am . 248, 120 .
PEREIRA, J. R., BUSTOS, R. E. and ZYNGIER, Z . (l%3) Some pharmacological properties of the juice of the
banana plant. Archs Int. Pharmacodyn. Ther. 144, 144 .
TEERAPONG, P. and HARVEY, A. L . (1977) The effect of S-hydroxytryptamine on the chick biventer cervicis
muscle . Eur. J. Pharmac. 44, 99.
TEERAPONG, P. and HARVEY, A. L. (1979) The actions of S-hydroxytryptamine on isolated skeletal muscle
preparations . Eur . J. Pharmac. 57, 329.
WMLKES, T. P ., S10ERDSMA, A., CREVELING, C. R. WEISSBACH, H . and IJDE :~FAIE~D, $ . (19Sß) Serotonin,
norepinephrine, and related compounds in bananas. Sclenet 127, 648 .
WEST, G . B. (t95ß) Tryptamines in edible fruits . J. Pharm . Pharmac. 10, 389 .