lar ridges on epidermal cells, warty protuberances on non-glandular hairs, and surface views of glands in ... of the various trichomes of Cannabis and also.
Amer. J. Bot. 63(5):
578-591. 1976.
TRICHOMES OF CANNABIS SATIVA
L. (CANNABACEAE)'
P. Da,vlNnNolN AND Pernn B. KLunvrlN Department of Botany, University of Michigan, Ann Arbor 48104 ABSTR A C T The diversity of non-glandular and glandular hairs of Cannabis sativa I:. (marihuana) are described by scanning electron microscopy. The non-glandular hairs are of two major types, as distinguished by size differences and locations, and all of them are highly silicified. The presence of silica as well as cystoliths of calcium carbonate help in the identification of marihuana even in its ash residues. X-ray microanalyses of Cannabis hairs are compared with those of Humulus lupulus and Lantana camera, whose hairs have been considered to resemble those of marihuana. Glandular hairs are found to be of two major categories, One group consists of glands whose heads are generally made up of eight cells and the other group whose heads are generally made up of two cells but never more than four cells. All glands of both categories are stalked. Some glands of the first category are massively stalked and these are restricted solely to anthers and bracts of staminate and pistillate plants. The massive stalk is considered to be made up of epidermal and hypodermal cells that have grown in response to some stimulation during anthesis. Fine details of the shoot system of Cannabis, such as cuticular ridges on epidermal cells, warty protuberances on non-glandular hairs, and surface views of glands in developing stages are also reported. Glandular hairs on the bracts of Humulus Iupulus resemblethose of Cannabis.
Mosr or rHE AERIALparts of Cannabissativa L. (the Indian hemp plant or marihuana) are covered with different types of non-glandular and glandular trichomes. In spite of its importance as an economic plant of antiquity, the bbtany of Cannabis is still poorly understood (Schultes, 1970; Schulteset al., 1974). This applies particularly to the nature of glandular hairs which are consideredto be the primary locations of the narcotic cannabinoids.Briosi and Tognini (1894, 1897) presenteda detailed description and illustrations of the various trichomes of Cannabis and also summarizedprevious works on Cannabis sative. BesidesMohan Ram and Nath's U964) cursory referenceto the ontogenyof the trichomes, most 'Received for publication 7 April 1975. The authors wish to thank Dr. Erich Steiner, Director, University of Michigan Matthaei Botanical Gardens for the supply of specimens,and Dr. W. C. Bigelow and his staff for use of the scanning electron microscope facilities. We also thank Diane Hoefle, Frederick Hebard, David Bay, and William Stein for their encouragement and assistance. This work was supported in part by NSF grant BMS 75-16359.
workers have describedsome aspectsof the mature trichomes only (Ballard and Phar, 1915; Bouquet,1950; Shimomuraet al., 1967; Jackson and Snowdon,1968). Recently,the needfor the identification of marihuana in illicit trade and possessionhas resulted in several publications, forensicin nature, which mostlv describethe nonglandular cystolith hairs. Theie studieshave employedeitherthe light microscope(Asahinaet al., 1967; Shimomuraet al., 1967; Nakamura,1969; Nordal, 1970; Thornton and Nakamura, 1972; De Forest,Morton and Henderson.1974) or the scanning electron microscope (Bradford and Devaney, 1970; Devaney and Bradford, 1971; Siegesmundand Hunter, 1971; Mitosinka, Thornton and Hayes, 1972). Such studieson the identification of marihuana have been carried out on resin, dried leavesor ash residuesof Cannabisas well as on the dried leavesof nearly 90 different plants that are said to possesscystolith hairs thdt could be confusedwith those of Cannabis. Hammond and Mahlberg (1973) made use of the scanningelectronmicroscopeto describethree
Fig. l-7. Surface features of regions of shoot system of Cannabis sativa. All figures obtained from fresh, live specimens without any treatment. l. A lobe of a young leaflet with reduced stomata near the tip, and several nonglandular trichomes. X 265, 2. Adaxial surface of a young leaf with cystolith trichomes of two sizes and a few bulbous glands. Some trichome initials are seen in the groove. y 260. 3. Mature region of a leaflet tip showing cuticular ridges and reduced stomata. X 265. 4. Surface of petiole showing non-glandular hairs, several capitate glands and a few bulbous glands. X 85. 5. Abaxial leaf surface with trichomes and stomata. Hairs are devoid of warty projections. Three capitate glands are seen to the left and two bulbous glands to the right. X 250. 6. Adaxial leaf surface. Cuticular ridges on the epidermis are yet to develop. X 250, 7. Adaxial leaf surface showing a capitate gland, long and short cystolith trichomes, and a ring of cells surrounding the long hair. X 250.
578
May*June, 19761
F'
w ri
r$$ h
DAYANANDAN
AND
KAUFMAN-TRICHOMES
OF CANNABIS
579
580
AMERICAN
JOURNAL
OF BOTANY
[Vol. 63
May-June, 19761
DAYANANDAN
AND KAUFMAN-TRICHOMES
types of glandularhairs that occur on the pistillate plants, especiallyon the flowering bracts. In this paper, we extend this study to the shoots of staminate plants which bear as many different types of glandular hairs as the pistillate plants. Using an X-ray detecting systemthat is coupled to the scanningelectronmicroscope,we also analysethe pattern of silicification and calcificationin the nonglandular hairs of Cannabissativa, Humulus lupulus (hops, the only other genus in Cannabaceae) and Lantana camera (Verbenaceae).X-ray microanalysisdata on the above plants, as well as on burnt ash residuesof marihuana, provide information of taxonomicvalue that make identification of marihuanamore definitive. Becauseof the variation that exists amongthe trichomes of Cqnnabis, their study should lead to a better understandingof the real taxonomic relationsthat exist betweenwidely different forms of Cannabis that have usually been grouped under one species,C. sativa (Schulteset al., 1.974). Marihuana contains about 20 different cannabinoids,including tetrahydrocannabinol,the major hallucinogeniccomponent. The study of Cannnbisglandsis also essential to understandtho biogenesis,distribution and function of the different cannabinoids. Mlrsnrars ANDMErHoos-A pistillate and a staminate plant of Cunnabis which have been growing in the University of Michigan Matthaei Botanical Gardenswere used for most of the investigation. A few seedlingsfrom the Botanical Gardenswere also employJd to stucly the differentiation of trichomes during early stages of growth of Cannabis. The plants employedfit the usual description of typical Cannabis sativa (Schulteset al., 1974) . Most of the scanning electron microscopic (SEM) studieswere done with fresh specimens. Small segmentswere mounted on alumirium stubs with carbon paint and examined imrrediately in the SEM (JEOL model JSM-U3); At an accelerating voltage of 15 KeV, fresh specimens could be examinedfor about 20 min without muclt damageto them. X-ray microanalyseswere also done mostly on fresh specimensby coupling to the SEM an X-ray detector and amplifier system (KEVEX) and a multichannel analyser (Northern Scientific,Model 710). Setected'areas of the specimenwere first photographedin the secondary emission mode followed by X-ray mapping for calcium and/or silicon. This order was nec-
OF CANNABIS
581
beam or vacuum_damage to the essarybecause. specimenwas increasedwith the longer exposure times that were used for X-rav analvsis. Samples were also subject'edto- critical-point drying in order to compare the structures with fresh sample preparations. Materials were fixed in FAA, dehydratedin ethanol and critical-point dried with liquid COz as an intermediate. Preparations thus obtained were as sood as those that employed amyl acetate insteid of ethanol. Critical-point dried samplesof Cannabis,as well as air dried samples of Humulus lupulus and Lantana camera (both obtained from the Matthaei Botanical Gardens) were coatedwith a thin layer of gold before examiningthem in the SEM. Ash residueswere obtained frorn dried leaves and flowering tops of pistillate and staminate plants of Cannabis. A marihuana cigarette was attachedto an aspirator through glassand rubber tubings and the ash falling from the lit end was collectedon clean piecesof aluminum foil. Small quantities of the finest ash sampleswere spread on scotch tapes and mounted on aluminum stubs to be gold-coatedand examinedwiththe SEM. Fresh samples of Cannnbis leaves that show minimal damageduring examinationwith the SEM are better suited for X-rav microanalvsis than dried and gold-coatedsampl-es.Howevei, in Canrwbis, Humulus, Lantana, and ash residues of Cannabis,silica and calcium deposition occur in such large quantities that no interferenceis encountered even with gold-coated samplesduring qualitative X-ray analysis for these elements. While this study focuseson SEM aspectsof the trichomes of Cqnnqbis, conventional microtechnical methods (microtome sections, acid-hydrolysed squashpreparations)were also employed to corroboratefindings obtainedwith the SEM. Os snnverroNs-Non-g IanduIqr hair s-The stems,petioles,stipules,leaf blades,bracts, and both surfacesof the tepals are all covered with hairs that are strictly unicellular (Fig. 1-7, 30, 31). Figure 1 depictsa young leaf which shows stagesin the developmentof hairs. The tip of a leafletpossesses 6-15 stomata,probablymodified to function as hydathodes. Immediately below the tip are well-developed and elongating trichomes. Warty protuberancescharacteristic of thesetrichomesare alreadyvisible (Fig. 1-3). At a still lower level, hair initials are seen among fully differentiated hairs. Such an admixture of
Fig. 8-13. Secondary and X-ray images of silicon of non-glandular hairs of Cannabis.-Fig. 8-11. Fresh, live specimens without any treatment. B, 9. Abaxial leaf surface. Note silicon localization only on hairs and also on a Two kinds of trichomes from adaxial leaf surface. Silicon contaminant, perhaps a soil particle. X 500. lO, ll. is localized in both types. X 520.-Fig. 12, l3,Trichome macerated in conc. nitric and chromic acid mixture, air dried and gold-coated. 12, 13. Isolated hair from the adaxial leaf surface. The cystolith of CaCO, has been removed but the region where it was attached still remains. Also seen is a broken piece of another hair. 1 950.
582
AMERICAN
JOURNAL
lVol. 63
OF BOTANY
HgEgE# u
E tf,' l =
EAEEg E €H
G
d
-u
O j Yi
o
-9*n -t =
oxHifre OE :S -P a
€ ^9 b
o>'.!:6Xi
s!
.'..i
;
i
SE€S#;
ctr6-'"i
E E c 6H E
!E3EgS N
OE
O
i !!G=-
#ooq8",:
Z 5E= EE *o,!?v:! !'
aE7
7
tsftcc
-
€t ia3?I .^r^
P .E d 3 r.P cjll 5 o E n.9d-H H 5F.-A t i
;:
.l :
=
Z ':i :1.::' : -' d q "'f! E .!
.a
Fl 6t
' -?-;crP i Fr-r f;*x;i E
^l ,E
t5 -* C.!Er X 6 bO o x-.;o I ifi
!
!
!
i " o**** ;€€ .--d oE € E Yi > >'x -
- -.. .P F i :-.G.l ;
;EE:F? o
3 g.9 s.5 icd>uE 5.- ::P E * o*l .!
;: P ! 'l
F
"
>q v
* :.e= i X X: H ='d Vi ;.N
ry.'lE 9 -E **
o q X o E 6 EA
-v I' boo 3t;
ll g T T e 9 " 9 . F-
A.=
?a HE dah
(D 'i 6i ! Y:
YE
,
E 9{.:'" ih 6: N ?.= -d.= a)1d.:31' !"ddIv
3- E;6Yb J*t e*;E sR bbA ui' iI 9e
.E Fc.$
iii |
SF-t Xd
.do
o#
Yl+
.-reS H .Ite
ir xEAEE
V
*
n
U:
.=\'=r o l e
o
rv
r
. 'j !
.9!-"HN*
,r
H
e o\ 5.q " H.* ;Erii C\J
May-June, 19761
DAYANANDAN
AND
KAUFMAN-TRICHOMES
583
OF CANNABIS
bs r#f EF
: 9 E UB : E n= XE PE'E
E .E;5 E i E o>i 9F - o*
tsFXlj-
trc
6 f ia . s F lE
is -i3E?E ";.Z!-.F A':.= ? E:i: ! i x 9 s Bs : . E . , 3 R! = . E 9 C -= I3' ;.:
' E3€:*EE !b Sa Eg
E EigEF: "
EEFETX=
E:EEfr:t OC
'l i
o
Y
eg;p::=H
;
@? i ,a
.9 6 u i=
6
s ff.El;8.
4:rEi€s
B gt .*; . E =. : 'd 9E"Y > l ; ! 'i = E o f l
SixE !:x 2.!a H Y.y. 3&4c'.2-
o o | 5'E -6':
Z HF T Ei F *Vxu Ia c E= ;€ "Er a iE; ;:3q^ff
='3.4k !.c ts; j:1Be *.E.9= E 9 F.E o'oE x
J g E € H frJ;
3 :;;; I i E rEENbX.96 fi +sE eE€ -q: d.E
. . t s : 'F a
C 'dF. !2S dE i gN E E c
ib€
Ed-
Jo
s g Ei ^gEE tE r.:E ;EH '*'=E E
x*'E
n t.E ' :
.P o.€
€ ::te ,i E @H.,!iF
/T v *:
h ci I ts ,, -!:O.:
.E
q!E; f 'EtN B
I r€isE^E ii=EEr:F d-i :9='FS U
** i i S s ,y -rj F g? F5 r *
AMERICA\
584
J OIIR \AI
OF BOT A\Y
lVol. 63
w
*j
,# @ ,,,
Fig. 30-33. Fresh, live and untreated specimensof male flowers of Cannabis. 30. Central region of a male flower showing the bases of stamens. No pistillode is seen. Glandular ancl non-glandular hairs are visible. The lattel have warty projections. The perfect spheresare pollen grains. X 180. 31. Basal region of outer tepal surface. Typical hairs and capitate glands are found, but the epidermal cells lack cuticulal ridges. A11 hairs contain Si. X 290. 32. Surface of anther towards the interior of flower with capitate glands arranged in a row. X 100. 33. Enlarged view of a capitate gland with a long stalk. Also note the stomata and cuticular ridges on epidermis. x 400.
May-June, 19761
DAYANANDAN
AND
KAUFMAN_TRICHOMES
well-developed hairs and hair initials in early stagesof developmentare also seenon young portions of stems,petiole basesand regionsof the leaf between the petiole and the bhdd 1Fig. 2, 34). The leaflet tips, and the small epidermalcells below them, are smooth and lack the characteristic ridgesthat developlater (Fig. 3). Even though glandular and non-glandular hairs are found at this stage,these regionsstill lack stomata (Fig. 1). As apicalgrowih of the leafletceases,an'dis stomatabegin to differentiate,the epidermal cells ellarge, and the leaflet tips and the adaxial epidermal cells develop the characteristiccuticular ridges (Fig. 3,23). This pattern is particularly striking on the leaflettips (Fig. 3). Two major 'of. types of non-glandularhairs cover the shoots Calnnobis.Th; shorter (about 70125pm) and prominent cystolith-containinghairs are generallyrestrictedto the adaxial leaf surfaces (Fig. 7). The longer (250-370pm) and more profuse hairs occur on the abaxial leaf surfaces, stems, petioles and tepals (Fig. 5, 8). All the hairs are pointed towards the tips of the stemsor leaves and exhibit varying degreesof warty protuberances. The cystolith hairs on the adaxial leaf surfacesexhibit differencesin size and wartiness (Fig. 6,7). Among the shorter (70r,m) ones,somenever developthe warty protuberances but they still containcystoliths(Fig. 10). X-ray microanalysisrevealsthat hairs from any part of the plant possesssilica (SiOz .nHrOi. Silica is distributed more or less uniformly all over the surfaceof the trichome(Fig. 8-13 ) . Deposition of calcium as CaCO3is confinedto the enlargedbasalportionsof the cystolithhairs (Fig. l4-I7). The cystoliths are very prominent in hairs on the adaxialleaf surfaces.Some,but not all the hairs from the stem, petiole, and abaxial leaf surfacesalso show cystoliths of CaCOs. Becauseno particular concentrationof any element is to be found on the warty prbtuberancesof the trichomes, the protuberancesare perhaps.due to local depositionsof cellulose,cutin or other wall materials. Remainsof marihuanaash revealthe existence of both kinds of hairs as recosnizableand distinct entities (Fig. 14-16). The cystolithsof CaCOs often separateout as balls. X-ray mapping of the ash further confirms the siliceous nature of the hairs and the presenceof Ca in the cystoliths (Fig. 16-18). Some Ca is probably distributed throughout the inner cavity of the hairs. But the enormous quantities of silica found on the cell walls of the hairs mask the small quantitiesof Ca (Fig.17). A study of the hairs of Humulus lupulus reveals that some are roughly comparablewith the two kinds of hairsfound in Cannabis@ig. 22,28, 29,42,43). However,the followingfeaturesreadily distinguishCannabis from hops. The abaxial leaf surfaceof Humulus is totally devoid of hairs
OF CANNABIS
585
exceptover the midribs and major veins (Fig. 38). The epidermis reveals intricate cuticular ridges which are not found on the abaxial leaf surfacein Cannqbis(Fig.5). The stem,inflorescence stalks, petioles and major leaf vein surfaces possess anvil-shapedunicellular hairs in hops (Fig. 39, 40). Hairs of this shapeare neverfound inCannabis. X-ray microanalysis data also bring out important differences. The hairs on the adaxial leaf surfacesof Humulus are poorly silicified. In contrast,a ring of epidermalcells around the base of each hair is highly silicified (Fig. 2a-27). Such basal rings of silicified cells are also found associatedwith the anvil-shapedhairs (Fig. 40,
4r).
Lantqna cqmera is said to possesshairs that are somewhatsimilar to that of Cannabis. The hairs of this member of Verbenaceaehave been described along with those of 90 other plants as being of forensic interest in the identification of illicit marihuana (Nakamura, 1969; Thornton and Nakamura, 1972). Viewed through the SEM, the upper leaf surface of Lantana only superficially resemblesthat of Canrwbis, La:ntana leaf epidermal cells lack the cuticular ridges and the hairs do not show warty protuberances (Fig. 19) . Lantana has prominentrings of basal cells. These cells are silicified while the hair proper is calcified (Fig. I9-2L). Glandulqr hairs-Glandular hairs also develop along with the non-glandularhairs. The types of glandular hairs that cover the aerial parts of the pistillate plants, especially those on the bracts, have been describedby Hammond and Mahlberg (1973). The latter authorsmadethe first attempt to classifythese glands as bulbous, capitate-sessile, and capitate-stalked. Briosi and Tognini (1897) describedthe bulbousand capitate-sessile glands of Hammond and Mahlberg as almost sessile, and the capitate-stalkedglands as stalked. Our studiesindicate that the diversitv of the elandular hairs has never been fully understoodand properly classified. Pistillate plants possessthe so-calledcapitate-stalkedglands only in the flowering bracts. Glands that could be described as capitate-stalkedalso occur on the staminateplants; these are restricted solely to longitudinal t'ows alongthe inner surfacesof anthers(Fig. 32,33). Briosi and Tognini (1897) describedthe stalks of the capitate-stalkedglands of the pistillate plants as primarily derived from epidermal and subepidermalcells at the bases of the glands. These glandspossessglobular heads and multicellular stalks that can attain a lensth of about 200pm. Some glands have much slorter stalks while othersappearto be almostsessile(Fig. 32). The importance of this gradation in stalk length, and the true nature of the multicellular stalk, are describedbelow. Glands with large globular heads, and apparentlv without stalks. have been described as
AMERICAN
JOURNAL
OF BOTANY
lVol. 63
May-June, 19761
DAYANANDAN
AND
KAUFMAN_TRICHOMES
capitate-sessile by Hammond and Mahlberg (1973). These glands are the most abundant type found, and they occur in especially high numberson the abaxial leaf surfaces,petiolesand young stems (Fig. 4-7, 34). The developmentof these glands from epidermal initials to the fully mature glands with a head of eight cells occurs rapidly during early stagesof shoot growth (Fig. 34, 56-62). These glands are not sessile. Each gland possesses a stalk that is only one cell in length but 2-4 cells thick (Fig. 35). In fact, the previously described capitate-stalkedglands are similar in all respectsto the so-called capitatesessileglandsexceptfor the presenceof a massive stalk. The massivestalk, however, is actually a product of the elongationof hypodermalcells below the glands. The cells that make up the heads of the glands of both types form a disc-shaped unit. fn early stagesthis cellular disc is convex and at maturity it may become either a flat or a concave disc (Fig. 36, 58-62). The disc is about 30r,rmin diam and about 15um in heisht. Because bt the accumulation of resinors se6retions betweenthe outer surfaceof the disc and the extended cuticular membrane, each gland gives the appearanceof a sphericalstructure(Fi9.34, 45, 60). Some of the FAA-fixed and criticalpoint dried specimensshowedsomeglandswhose outer membraneshad ruptured, thus revealingthe inner cellulardisc (Fis. 36). Small glandular trTchomes(about 15-30pm) describedas bulbous slandular hairs are considered by Hammond anl Mahlberg (1973) to be least complex structurally. Our findings indicate that these smaller glands are quite variable (Fig. 46-50). Somehaveheadsmadeup of singlecells and they have stalks that are unicellular. Other glands have four-celled headswith stalks that are two cells thick. The majority of the small glands possesstwo-celledheadsand.stalksthat are one or two cells long and one or two cells thick (Fig. 51-55). This variability was observedamong mature glands. Unlike the capitate glandsof sev-
OF CANNABIS
587
eral-celledheads,the smaller glandsoften possess chloroplastsin the stalk as well as in the head cells (Fig. 51-55). Resinoussecretionsare found to accumulatewithin the cells of these smaller glands. With maturity the resin collects beneath the outer membranesof the glands,giving the appearanceof nipple or flask-shapedunits (Fig, 47, 48, 53). That theseare not artifactsinducedby the vacuum in the SEM was evidencedby the observation of such features in fresh specimensexamined with a light microscope. Figure 40 shows a region of the outer membranethat is different from the rest of the cuticular membrane. When young leaves are hydrolysed in 1N HCI and squashedbetween a slide and cover glass, this resion is detached and comes off as a contact leis-shapeddisc (Fig. 48,52). DrscussloN-Although the descriptionthat follows and all but one fizure in this article are concerned with the stamlnate plant, we found the same types of glandular and non-glandularhairs on the pistillate plants also. The exceptionsin distributions are to be found only among the male and femaleflowersthemselves. Both glandular and non-glandulartrichomes of Cannabiscover the shootsystemof Cqwwbis fuom the early seedling stagesto maturity. Massively stalked glands (capitate-stalked), however, are restrictedto the bracts of the pistillate plants and anthers of the staminateplants. The assumption that resinousglandsdo not developuntil the plant is on the point of flowering is incorrect (Bouquet, 1950). Staminateplants do not possessa restricted number of glandular types as stated by Hammondand Mahlberg (1973). Both the sexes bear similar types of glandular and non-glandular hairs. Non-glandular trichomes that cover the plants are of two major types. The predominantly cystolith-containins hairs are found on the adaxial surfaceof the liaves. Differencesin size, as well on them, as the existenceof warty protuberances
