Document 06OJaEaQGmaEER9mZjE2n3wM

lographic traces of compound action potentials obtained under experimental conditions identical to those of the up* per portion of Fig. 2, except that the ganglion had been superfused for 24 minutes with 5 vnM theophylline. The unconditioned test response (trace c) and the conditioning response (early part of trace d) were unaffected by theophylline. In contrast, the condi* tioned test response (later part of trace d) was considerably more reduced in amplitude (46 percent) and area (44 percent) than prior to theophylline. This increased inhibition of synaptic transmission after a conditioning stimu* lus, observed in the presence of the ophylline, can be attributed to the potentiation of the s!ow*lPSP achieved by this phosphodiesterase inhibitor (Fig. 1A). Our results indicate that cyclic AMP can mimic the electrophysiologic effects of dopamine, a putative ganglionic ncurotransmittcr. Cyclic AMP has been found to mimic tho hypcrpolarizing action of /3-adrencrgic agonists on the Purkinjc cells of the rat cerebellum (12) and on the smooth muscle cells of the rabbit pulmonary artery (13). In the ease of the Purkinjc cells, theophylline potentiated, and PCEt blocked tho inhibition of spontaneous discharge caused by application of exogenous norepinephrine (J2). Our electrophysiological data sup port the hypothesis that cyclic AMP plays a role In synaptic transmission in sympathetic ganglia. There appear to be both direct excitatory and inter neuron-mediated inhibitory input from the preganglionic fibers to the post ganglionic neurons of the superior cer vical ganglion (10, 14), Our evidence supports the idea (2-4) that the slowIPSP is generated by an increase in the amount of cyclic AMP in the post ganglionic neurons in response to dopamine, released from the interneu rons. The hypcrpolarization of the post ganglionic neurons makes them less responsive to subsequent excitatory in put. According to this scheme, cyclic AMP mediates dopaminergic transmis sion and, thereby, modulates choliner gic transmission in the ganglion, the modulation being of an inhibitory type thnt produces a negative feedback and limits the effectiveness of subsequent excitation. Donald A. McApeb Paul Greengard Department of Pharmacology, Yale University School of Medicine, New Haven, Connecticut 06510 311 Rtlcmcti Ml Note* 1. D, A. McAfse, M, Schorderet, P. Grtengsrd, Pharmacologist 12, 488 (1970). 2. --------- , Science 171, 1156 (1971). 3. J. W. Kebablan and P. Greengard, Ibid. 174, 1346 (1970. 4. P. Greengard, D. A. McAfee, J. Kebablan, in Advances In Cyclic Nucleotide Research, P. Greengard, R. Paoletii, G. A. Robison, Eds. (Raven, New York. In press), vol. I. 5. B. Libel and T. Tosaka, Croc. Not. Acad. Scl. U.S.A. *7, 667 (1970). B. Libel, Fed. Proc. 29, 1945 <1970). 6. P, Greengard and D. A. McAfee, Blochem. Soc. Symp., in press. 7. D. A. McAfee, P. Kells, M. Schorderct, P. Greengard, In preparation. 8. H. W. Kosicrllte, G. M. Lees. D. I. Wallis, J. Physiol. London 195, 39 (1968). Ganglia were removed from New Zealand white rab bits under urethane anesthesia (I g/kg), were decapsulated, and were positioned In the su crose gap apparatus. The preganglionic (cervi cal sympathetic) nerve was stimulated through platinum bipolar electrodes. Locke lolution containing various agents superfused the gan glion M ) It 2 ml/mln. A short segment of ihe proximal portion of the postganglionic (Internal carotid) nerve was superfused with 315 mAf sucrose, < 0.3 mt/min, and normal Locke solution superfused the distal end of the postganglionic nerve at 0.5 to 1 ml/mln. Solution flow through the various chamber* was monitored by precision Row meters. The d*c potential difference between the ganglion and the end of the postganglionic nerve (that Is. across the sucrose gap) was amplified and displayed on an oscilloscope and on a stripchart servo recorder. The potential difference across the sucrose gap with the ganglion in normal Locke solution was 5 mv or taas, and the resistance was usually 0.4 to 0.1 megohm. The design of the sucrote gap ap paratus and the precise control of flow rates reduced flow artifacts and eliminated as sociated problems. The composition of the Locke solution (In millimoles per liter) wea: NaCI, 136; KCI, 5.6; NaHCO,, 20.0; NaH,PO,, 1.2; CaCt* 2.2; MgClr 1.2; and glucose. 5.5. The Locke solution was equili brated with a gas mixture of 95 percent O end S percent CO, and had a pH of 7.2 to 7 3 at the temperature of the experiment* 23* to 25*C). 9. W. ). George, J. B. Poison, A. O. O'Toole. N. D. Goldberg, Proe. Nat. Acad. Set. V.S.A. 66, 398 (1970); J. A. Ferrendelll, A. L. Steiner, D. R. McDougal, D. M. Klpnls, Blochem. Blophys. Res. Commun. 41, 1061 (1970); I. F. Kuo. T.-P. Lee. P, L. Reyes. K. G. Walton, T. E. Donnelly, P. Greengard, /. Biol, Chtm. >47, 16 (1972). 10. R. M. Ecdrs and B. Libel, i. Physiol. London 157, 484 (1961). 11. A. P. Sumlyo, A. V. Stimlyo, N. Friedman, Ann. N.Y. Acad. Scl. 185, 101 (1971). 12. G. R. Sigglnt. B. I. Hotter, P. E. Bloom. Science 165, 1011 (1969). 13. A. V. Somlyo, O. Haeusler, A. P. Somlyo, Ibid. 1(9, 490 (1970). 14. T. H. Williams and S. L. Palsy, Brain Res. 15, 17 (1969); A. Bltlrkiund, L. Ctgrell. B. Falck, M. Rdiln. E. Rosengrcn. Acta Physiol. Scand. 78. 3)4 (1970). 15. Supported by PHS grants N3 00440 and Mil 17387, NSF gram OB 27510, anti a postdoctoral fellowship from the Connecticut Heart Association to D.A.M. 19 May 1972: revised 21 July 1972 g / ! r~ ''-'.r..,., t:. <*' 1 Polychlorinated Diphenyls: Melabolic Behavior of Pure Isomers in Pigeons, Rats, and Brook Tront Abstract. The metabolic behavior of pure mono-, di-, tetra-, and hexachlorobiphenyl homers in pigeons, rats, and brook trout was investigated. Excreta from these animals were extracted and examined by chromatographic and mass spectromctric techniques. The results showed conversion of the 4-chloro-, 4,4'-dlchloro-, and 2J',5,5'-tetrachlorobiphcnyi homers into monohydroxylated de rivatives by the rat and pigeon whereas no hydroxymetabolites were detected in the excreta of the brook trout. No hydroxylated products of 2J',4,4'JJ'hexachloroblphenyl were detected in the excreta of pigeons, rats, or brook trout. Polychlorinated biphenyjs (PCB) are row recognized as almost universally distributed pollutants which are gener- ally considered to be quite resistant to chemical and enzymatic degradation. Recent evidence suggests that chloro- Table 1. Data on hydroxylaled metabolites from rat urine and piieon excreta. Thin-layer chromatography on silica; solvent A, hexnnc-aceionc, 2.5 : 1; solvent B, benzene-ethyl acetate. n : 1. Mydroxychlorobipheny) Ry in solvcntt Mass spectrum Compound administered* Number of A B Molecular ion chlorine atom* in the metabolite 4-Chlorobiphenyl 4-Chlorobjphenyl 4,4'-Dichlorobiphenyl 2,2',5,3-Tetrachlorobiphenylt 0.3 0.5 204 1 <0.3 220 It 0.5R 0.6 23811 2 0.535 0.55 306 4 * No hydroxymetabollles could be delected in the excrete of rats or pigeons treated with 2,2',4,4\5,5`. hexachloroblphenyl. t The PCB isomers move with the solvent from In these systems. I This compound was found In rat urine only (that Is, not In pigeon excreta). | Several minor bands In this region were extracted together; peaks due to impurities were present In the mess spectrum at different temperatures. Compounds were made visible by viewing with ultraviolet light or by (praying a portion of the plate with a I percent soluilon or 2.4,7.irinltn>.9-fluorenone in acetone. 1 The accurate mass was determined for this compound. Mass calculated for C.JI.CLO: 237.9952; mast found. 237.99S9. WlENCr. VOl |7n MQN5 084563 approximately 30 seconds. The slow- Conditioning ganglion in solutions of dibutyryl cyclic JPSP, but not Ihc inilia] EPSP, was GMP for longer periods of time (5 lo potentiated in amplitude and duration 10 minutes). Higher doses (100 to in she presence of theophylline (Fig. 250 fiM) of dibutyryl cyclic GMP I A). Theophylline did not inhibit either greatly enhanced the rate of depolariza the amplitude or the duration of the tion of the postganglionic neurons but slow-EPSP (not shown). The hyper* either had no effect on, or decreased pollination of the postganglionic neu the size of. the transient hyperpolurizu- rons induced by exogenous dopamine (ion. These observations arc in contrast was also potentiated by theophylline to those made on liver slices where (Fig. 1C). The response to dopamine was tested by switching for 3 minutes from a superfusate of normal Locke solution to one containing 50 dopa mine (prepared 2 to 5 minutes previ ously) and then by returning to the normal Locke solution. Theophylline (I to 5 mM) in nine experiments caused an increase of 44 7 percent (mean S.12.M.) in the amplitude of the slow-IPSP, and an increase of 54 9 percent in the amplitude of the hyperpolaritation induced by 50 to 200 fiM dopamine. The effect of the theo phylline could be reversed by super fusion of the ganglion with normal Locke solution for 90 minutes. In each of 15 ganglion prepara tions studied, I0~TM to I0-0M PGEj virtually abolished the slow-IPSP within 10 to 20 minutes and substantially reduced the slow-EPSP, but had no Fig. 2. Compound action potentials, elicited by submaximum stimulation of the preganglionic (cervical sympathetic) nerve, and recorded from platinum bi polar electrodes on the postganglionic (internal carotid) nerve 2 to 3 mm distant from the pole of the ganglion. Upper row, superimposed traces a and b recorded during superfusion of the gan glion with normal Locke solution. Lower row, superimposed traces c and d re corded 24 minutes after start of super fusion with Locke solution containing 5 mM theophylline. Traces a and c, re sponse to a single test stimulus only. Traces b and d, response lo a condition ing stimulus followed 600 msec later by response to a test stimulus. Conditioning stimuli were just maximum for the S, elevation [see (/0)J: voltage of the test stimuli was one-half that of the condition ing stimuli; pulse width for both kinds of stimuli was 0.5 msec. Calibration mark: 200 it*. 100 msec. Bandwidth, 2 hz to 10 khz. both cyclic AMP and cyclic GMP caused a hypcrpolarization (11). Con ceivably, cyclic GMP may mediate the slow-EPSP and, thereby, increase the responsiveness of the postganglionic neurons to subsequent excitatory input. If so, this would indicate that cyclic AMP and cyclic GMP function in op posite directions, that is. in a push-pull fashion to exert long-term control over neuronal excitability in (he sympathetic ganglion. Dopamine, theophylline, PGEt, cyclic GMP, and cyclic AMP and its butyryl derivatives, when tested on axons of the cervical vagus nerve in the same concentrations that had been used on the ganglion, were found to cause little or no effect on the membrane potential. High concentrations of dibutyryl cyclic GMP (I to 4 x 10-<M) had an effect on the vagus nerve (Fig. 1G) similar efTcct on the initial EPSP. (A transient to that which had been observed oo hyperpolarization of the postganglionic permeability of neuronal membranes to the ganglion with lower concentrations. neurons of 5 to 10 minutes duration, was usually observed on starting POE] superfusion.) The effect of 3 x 10-TM POEt on the slow-IPSP is illustrated cyclic AMP. Studies have shown that acetylcholine can cause an increase in the amount of guanosine 3',5'-monophosphate (cy Wc have been able to demonstrate consistently and reproducibly, with each of 13 preparations, that the excitability of postganglionic neurons in the supe in Fig. IB. A concentration of POE] clic GMP) in heart and brain tissue rior cervical ganglion was diminished of 1 x I0~ "A/ caused a 50 percent (P). Moreover, available evidence indi during the period of the slow-IPSP and decrease in the amplitude of the slow- cates the existence, on postganglionic that this inhibition was markedly po IPSP. In nine ganglion preparations, neurons of the superior cervical gan tentiated by theophylline. These effects 1 the effect of POEt was tested on the hypcrpolarizatioo induced by 50 to 200 fiM dopamine. It was found that glion, of muscarinic-type reeeptors that respond to acetylcholine by causing a prolonged depolarization of the neurons are illustrated in Fig. 2 where the un conditioned response to a submaximum test stimulus is compared with the re PGK, abolished, or largely reduced, the (10). These same receptors are prob sponse to the same strength of stimulus dopamine-induced hyperpolarizatlon at ably involved in the generation of the applied 600 msec after a stronger con the same concentrations as were effec slow-EPSP (10). In view of the possi ditioning stimulus, that is, at the time tive In abolishing the slow-IPSP (Fig. bility that cyclic GMP might mediate of maximum development of the slow- ID) . The effects of PGEt could be this muscarinic depolarizing action of IPSP. The upper portion of Fig. 2 con largely reversed on prolonged super acetylcholine, we have studied the tains two superimposed oscillographic fusion witlt normal Locke solution. effect of cyclic GMP and dibutyryl traces of compound action potentials Monobutyryl cyclic AMP, applied in cyclic GMP on the resting membrane derived from the postganglionic nerve a concentration of 1 to 2.5 mM, hyper- potential of postganglionic neurons. with platinum bipolar recording elec polarized 8 of 11 ganglia tested (Fig. Cyclic GMP itself did not cause a trodes. The test response in trace a IE) . Qualitatively similar results were change in membrane potential. How was elicited without a prior condition obtained with cyclic AMP and dibutyryl ever, exposure of the ganglia to dibu- ing stimulus and therefore did not cyclic AMP, but not with adenosine lyryl cyclic GMP, in low concentrations occur during a slow-IPSP. The test 5'-monophosphatc, adenosine, or bu (2.5 lo 5.0 X 10"nM) for 4 minutes, response elicited 600 msec after a con tyric add. Cyclic AMP and its deriva caused a small, transient hyperpolariza ditioning response (trace b) and, there tives were never observed to cause tion followed by a depolarization of the fore, during the slow-IPSP, was reduced depolarization. The lack of any re postganglionic nerve ceils in each of 24 percent in amplitude and 20 percent sponse, by some of the preparations, seven preparations tested (Fig. IF). A in area compared to the unconditioned to the direct application of cyclic AMP depolarization of several millivolts test response. The lower portion of Fig. or its derivatives may be due to a low could be achieved by maintaining the 2 contains two superimposed oscil- 20 OCTOBER 1972 HONS 084564 Ml mr biphenyls do show chemical changes when exposed to ultraviolet light (/). Experiments with commercial PCB preparations also indicate that certain components, mainly those of lower | Pigaon chlorine content, arc metabolized by rats (2), pigeons (3, 4), and Japanese meUbotites oo- <DrQ-ci OiO-" quail (.?>. The evidence is based on the disappearance of certain peaks in the gas chromatograms of PCB ex truded from tissue. A similar disap pearance of peaks, on the other hand, was not observed in fish fed food con c,hO-Q-ci taining Aroclor 1254 (5). Because of their complex isomer composition, commercial PCB prepara tions arc unsuitable for a study of the metabolic behavior in which individual compounds (metabolites) arc to be Cl Cl Ral Pigeon CM Cl Cl identified. We know of only one study of the metabolism of a chemically de fined chlorobiphcnyl in (he literature (6). In this case 4-chlorobiphenyl was shown to be converted to 4-chloro-4'- No hydroxyialad malabolite* Ral Pigeon No hydfoiyli|| metabolites hydroxybiphenyl and its conjugates by the rabbit. We have examined the metabolic be Fig. I. Scheme for (he metabolism of chtorohiphenyls in three animal species. havior of 4-chlorobiphcnyl, 4,4'-dichlorobiphcnyl, 2,2',5,5'-tc1rachIorobipltenyl, and 2,2',4,4',5,5'-hcxachlorobiphcnyl in laboratory rats, Carncau pi Each brook trout was fed a single dose of the chlorobiphenyl (460 mg to 1.2 g/kg) (the compound was con cealed in chunks of beef liver). Im could be detected in the urine of rata treated with 2,2',4,4/,5t5''hexachlorobiphcnyl but the unchanged compound was found in the feces. In all instances geons, and brook trout in an effort to mediately after being fed, the trout no evidence was found for reductive observe (I) species differences and (ii) were transferred to a 5-liter tank and dechlorination in the animals tested. differences in the metabolic behavior of kept there for 4 days. The water io The metabolism of the four chlorobi PCB components of different chlorine which the fish were kept during the 4 phenyls in pigeons appears to be similar content. The compounds were chosen days after treatment was concentrated to that in rats except that no dihy- for their occurrence in commercial to 1 liter, acidified with concentrated droxychlorobiphcnyl was present in ex N PCB preparations (7). hydrochloric acid, and refluxed for 3 creta from birds fed 4-chlorobiphenyl. An inlraperitoneal injection of the hours. The acidified solution was ex The spccifipity of the hydroxylation chlorobiphcnyl (9) dissolved in oil tracted with ether, and the extracts were reactions is not known. On the basis (50 mg/kg) was administered to each prepared and examined for hydroxy- of qualitative observations, the ease of of the young male rets housed in meta chlorobiphcnyls as described above. formation of hydroxychlorobiphenyls in bolic cages each day for 3 days. A Data for products obtained from rats and pigeons seems to be as follows: capsule containing a solution of the animals treated with chlorobiphenyls 4-chloro 2s 4,4'-dichloro > 2,2',5,5'- chlorobiphenyl in corn oil (15 to 20 are reported in Table 1 and Fig. 1. In tetrachloro > 2,2',4,4',5l5'-hexachloro. percent) (60 to 100 mg/kg) was fed agreement with earlier observations on For the conditions and methodology to each of the pigeons each day for the metabolism of chloronaphthalenes used here, there was no evidence for 3 days. Urine and feces were collected (9) and chlorobenzenes (10), mamma the excretion of hydroxylated chloro for 1 week from the start of the treat lian metabolism (hydroxylation) of biphenyls in brook trout. ment from treated and untreated rats. Urine samples were hydrolyzed by re fluxing with an equal volume of 8N sulfuric acid for 1 hour and extracted chlorobiphenyls seems to become in creasingly more difficult as the number of chlorine atoms in the molecule in creases. In the rat, 4-chlorobiphenyl is O. HuTziNor.R O. M. Nash S. Safe Atlantic Regional Laboratory, with ether. The feces were extracted metabolized to a mono- and a dihy- National Research Council of with hexane at 50*C. In the case of the droxychlorobiphenyl, and only little Canada, Halifax, Nova Scotia pigeons, the total excreta for 1 week starting material could be recovered from the start of the treatment were from the feces (no starting material A. S. W. DlFrvitas R. J. Noksirom hydrolyzed and extracted. By compar could be recovered from the urine). In Division of Biology, ing thin-layer chromatograms (Table the case of 4,4'-dichloro- and 2,2',5,5'- National Research Council of 1) of samples from the treated and tetrachlorobiphcnyl, large quantities of Canada, Ottawa, Ontario untreated animals, the extra bands due to the chlorobiphcnyls administered could be located, extracted, and ana lyzed by gas chromatography and mass spectrometry. the unchanged material could be ex tracted from rat feces and only a mono hydroxy derivative was identified in rat urine. With the method used in this experiment, no hydroxymctabolites D. J. Wildish V. ZlTKO Fisheries Service Biological Station, Environment Canada, St. Andrews, New Brunswick 2n orroim* 1972 313 HONS 0*4565 i m y - - h -m % f RiltnuN an# NMm 1. S, Safe and O. Mullingar, Nature U>, Ml (1971); K, Huaterl and P. Kf1, Chtmotphtre I, 7 (1972); O. Hutclnger, S. Safe, V. lllVn. Environ. HrmtiH frrjprcl, 1, IS (1972). 2. D. L. Oram, W. E. T. Phillip*, D. C. Villa, new*. Bull. Environ. Contain. Toxicol, 6, 1112 (1971). 3. S. Bailer and P. I. Runyan, Nature 2*. 94 (1972). 4. A. S. W. DeFreltai, X, T. Nonirom, O. HutUnger, paper preacmcd at the iHlh meeting of thr Amrrteal Chemical Socielr, New York, IV72 (paper No. 32, Dlvliion of Water, Air, and Watte Chemlatry). 5. V. Zliko and O. HutHngrr, In preparation. 6. W. D. Block and H. H. Comlth. /. Biol. Chtm. 04, 3301 (1939). 7. D. Slttonr and D. Welti, /. Ckromoiogr. M, 13 0071): R. 0. Webb and A. C. McCall, /. An, Og. Anal. Chtm. SS, 746 (1972). B. O. Huulnser, 5. Safe, V. Zilko, Bull. Ehvlron. Contain. Toxicol, d, 209 (1971); S. Safe and O. Hutzlnger, i- Chtm. Soc. Perkin Trout. I, 6S6 (1972). 9. K. H. Comlsh and W. D. Block. /. Biol, Chtm. S31, SI3 (1938). 10. D. V. Parke nd R, T. William*. Blothtm. J 74, S (I960), and reference* died (herein. 11. We thank Dr. D. J. Ecobichon who supplied umple* of rat urine and fecea and Dr. W. D. Jainieaen who carded out the hlgh-resolu* tlon maw ipectral measurement*. National Retearch Council of Canada contribution No. 12733. 16 June 1972 B Raphides with Barbs aiid Grooves In Xanthosoma sagittifoUum (Araceae) Abstract. Raphides In petioles of Xanthosoma sagittifolium are needlelike crystals about 50 micrometers long. The rectangular cross sections have maxi mum dimensions of approximately 850 by 250 nanometers. The raphides have two distinct end structures. One end is narrow, acute, and tapered to a point; the other is broad, acute, and abruptly pointed. Barbs, about 750 angstroms long with tips oriented away from the narrow end, occur along the length of i/if raphide on ridges on either side of two iongtiudtnal grooves. These grooves, located opposite each other, give the raphide cross section an H-shape. Raphidcs are needle-shaped crystals of calcium oxalate,. occurring in bun dles within specialized cells of certain flowering plants (/). Ingestion of fresh plant tissue containing raphides usually results in immediate and often severe irritation of the mouth and throat. Two reasons for this irritation have been suggested: (i) mechanical irritation by the crystal itself (2) or (ii) chemical irritation by a curare- like drug associated wkh the crystal (J). We describe barbs and grooves on raphides, which probably act as me chanical irritants and possibly act by carrying a chemical irritant into the wound produced by the crystal. To our knowledge, the barbs have not been reported before. Xanthosoma sagluifolium (L.) Schott, commonly known as xannia or yautia, is an edible aroid grown as Fig. I. (A and B) Light micrographs. (A) A specialized cell, containing many raphides in a bundle, protruding into the petiolar air canal, and (B) a single raphide. (C and D) Scanning electron micrographs. (C) Two raphides showing barbs and grooves, both at the broad, acute, abruptly pointed end. The broken one shows the two grooves on the narrow aides of the raphide. (D) The narrow, acute, tapering point of the raphide. Note the orientation of the tips of the barbs away from this end. a subsistence or commercial crop in many Pacific islands (4). The corm is baked or boiled and eaten as a source of siarch. Leaves of other species of Xanthosoma are cooked and eaten in the same way as spinach. Specialized cells containing raphides occur in all organs of these plants. In this study petiolar material was collected from plants grown at the Lyon Arboretum at the University of Hawaii. Fresh material was observed with a Zeiss RA light microscope. Sections of tissue for scanning electron microscopy were crushed on the sample holder to force mechanical re lease of the raphides from the spe cialized cells in which they developed (Fig. 1A). The samples were dried in air for S minutes, coated with an AuPd alloy in a vacuum evaporator, and viewed and photographed with a JEOL JSM-U3 electron microscope op erated at 15 kv. The use of crushed tissue in the scanning electron microscope is an excellent method for observing ra phides. Raphides are held in the tissue, and the irregularity of the tissue sur face makes it possible to observe the crystal morphology in various orienta tions. The needlelike raphides of X. sagluifolium are about 50 nm in length (Fig. IB). They have two dis tinct end structures. One end is broad, acute, and abruptly pointed (Fig. IC), and the other is narrow, acute, and tapering (Fig. ID). Two grooves lo cated opposite each other run the length of the crystal. The cross sections have maximum dimensions of approxi mately 850 by 250 nm. The grooves in the sides give the cross section an H-shape (Fig. IC). Similar cross sections can be seen in transmission electron micrographs of Lemna and Spirodella (5). No pre vious description of gioovcs has been made. In X. sagittifoUum the grooves are shallow near the narrow point, but they appear to extend almost to the tip (Fig. ID). These grooves may al low material (possibly a chemical irritant) to be carried into tissues with the raphide. The grooves may also prevent throat or mouth tissue from sealing around the raphide. Their small size may allow tissue fluids to leak along the groove, as occurs in the blood channels of some weapons. Prominent barbs occur on ridges on either side of the grooves along the length of the crystals. The tips of the barbs are oriented away from tho narrow end (Fig. ID) and toward the SCIENCE. VOL. 17* MONS 004566