THE ORIGIN OF ADVENTITIOUS ROOTS IN CUTTINGS OF PORTULACA OLERACEA L. MARY
H.
CONNARD AND
P. \iIl.
ZIMMERMAN
In stem cuttings the place where roots emerge varies with the species.
I t might be assumed, therefore, that some internal differences also occur.
This may be one reason why botanists are not entirely agreed as to the
exact internal origin of adventitious roots from stem cuttings_ In this paper
the authors will not discuss the general subject of the origin of these roots
but will set forth the situation as it appears in a specific case, namely in
cuttings of Portulaca oleracea L. For that reason a long review of the litera­
lure will not be made_ Il will suffice to refer to only two papers_ In 1888
Tieghem and DOllliot (2, pp. 409 and 488) concluded that endogenolls ad­
ventitious roots occurring naturally in young portions of stems originated
entirely in the pericycle. They state that in older portions of the stems, or
later in ontogeny, roots might arise in phloem parenchyma, or still later in
the cambium.
In 1929 Priestley and Swingle (I, p. 64), after reviewing the litera­
ture, wrote, "The points that should again be emphasized are the pericy­
dic origin of roots in young stems, and the origin in the neighborhood of
the cambium in older stems; in either case the roots are generally inti­
mately associated with the rays; furthermore, their origin involves more
than one layer of cells."
MATERIALS AND METHODS
The material used consisted of cuttings of purslane (Porf'uiaca oleracca)
collected from the gardens of Boyce Thompson Institute. The cutting;;.;
were made at various times from July to September and placed in water
or sand media 10 permit of root gro'Nth. Properly selected cuttings made in
July rooted uniformly in four to six days. The roots arose only at the base
of the cutting anu emerged through the cut surface parallel to the main
axis of the stem (Fig. I). To obtain sections with root primordia, basal
pieces of the stems about 5 mm. long were cut at intervals beginning 40
hours after the cutting was planted and killed in chrom-acetic solution.
These pieces were embedded in paraffin and then cut into'serial sections 12
to 18 j1 in thickness. Safranin and Delafield's haematoxylin were llsed to
stain most of the sections.
STEM STRUCTURE IN GENERAL
Before one c<tn lInderstand {he origin of auvcn1iLiolls roots of pllr~!a)ll',
it Ie:; necessary (0 kllo\,,," sOlllething about its general JllOrpbolngy. Thepllr~­
lane stem j" uf a conlmou herbaceuus type with uiscrete bunules. The pri­
CONTRIBUTIONS FROM B OYCE TH OMPSON INSTIT UTE
IVOL · 3
FIGURE 1. Cuttings of Portulaca oleracea showing roots emerging fro m the cut surfa ce.
T he epidermis had been removed from the basal part of A in a n attempt to induce the roots
to emerge t hrough the cortex instead of t he cut s urface.
193 11
C ONNARD AND ZIMMERM A N -
PORTULACA OLERA CEA ROOTS
339
mary vascular system con sists of fo ur or five main bundles which divide,
formin g traces for leaves and branches . The branch traces are formed a
s ho r t distan ce (about 2 mm. ) be low t he point where they leave the s tem so
th a t they a ppear only in section s near a node (PI. I , J). On th e o ther hand,
as soon as o ne set of leaf traces leaves the stem the ma in bundles start
bra nchin g to form those fo r t he leaf directl y a bove so that t he leaf traces
a re seen thro ug hout the intern odes. In a ny internode there a re present the
main bundl es alternating with gronpsof from one to three leaf traces mak­
in g from I O to 20 bundles in t he stem. As t he stem grows old er so me sec­
ondary tissues develop both from t he fasc icular and in terfascicul ar cam­
bium s. The ac tivity of the latter is la rgely confined to the format io n of
small bundles (PI. I, 4) which run m ore or less obliquely connecting t wo
hundles , ge nera lly a main bundle a nd a leaf tr ace.
The ste m is s urrounded by a sing le, tho ugh occasionally double, laye r
of epid e rma l cells. Periderm may OCC UI" thro ugh wounding but is no t no r­
m a ll y produced. Within the epide rmis a re two to three layers of collen­
chy ma cells wi th heavy cellulose wa lls. There a re fo ur to s ix more layers o f
cortical cells be fo re reachin g the endode rmis o r starch sheath. These cells
are inte rmediate in size between those o f t he cortex and the smalle r o nes
of th e ra y a nd are somewhat less spheri ca l t ha n eit her. No thickenin g of
their wall s has been observed. Occasionally the ceIls contain mOre sta rch
than the s urrounding tissue but it is often di ffic ult to distinguish this la yer
especia lly when distorted by root g row t h. I mmediately within th e en­
dodermis, o uts ide of the vascular bundles , are g roups of bast fibers. Th ese
and th e s in g le laye r of cells in the ray in line wit h t hem and adjacent t o th e
endodermis have been considered to constitute t he pericycle. Ou its inne r
side the pericyd e throughout mos t o f the ray adj o ins the in terfascicula r
cambium.
ADVENTITIOUS ROOTS
A s ing le c utting may produce mo re th a n a hundred roots but usually
the numbe r is neare r fifty. There is no appa rent regularity in arrangem ent
nor in order o f development. The roots generall y arise in the rays adjacent
to the prim a ry bundles. In older stems th ey ma y a lso appear adjacent t o
the smaller co nnecting bundles or completely ind ependent of any pre­
viously form ed vasc ul a r tissues (PI. I, 3 a nd 4). The root primordia a re
often located one a bove the other alon g the vasc ul a r bundles (PI. II, I ).
On the average , eac h bundle has associated with it two or more roots with­
in 2 to 3 mm. o f the base of the cutting . M ore roots a re apt to be asso­
ciated with the la rge r bundles than with the sma lle r ones. As man y as ten
have been observed which ha d started alon g o ne s ide of a main bundle . Of
these the uppe rmost was no more than 4.5 mOl . from the c ut s urface. When
roots arise very close a bo ve each other "twin" roots may occur with two
s teles.
340
CONTRIDUTI ONS F ROM BOYCE THOMPSON INSTITUTE
[VOL ·3
EXPLANATlON, PLATE I
A ll phot-oillicrograp hs a re X47 except 1 a nd 2, which are X {3. Lette rs designating cer­
taill structures have the foll owing meaning: e, endode rm is; b, bast; f), pericycle; c, cambjum .
1. Cross section a bove the r ootin g reg ion 2 .2 m m. fro m the base of the cutting. F our
of t he ma in bundl es have bran ched, forming b ranch traces. The ot he r smaller bundl es in
groups of three ace leaf traces.
2. Deta il of t he uplJe r gra UIJ of leaf traces of r.
3. Cross sect ion of a stem 0.4 mm. from the base, showing root s dissolving the cortex .
<1. Cl'OSS sectio n 0.5 5 mm. from base. showin g the sa me lea f traces as 2. The root s a t
th e left a nd right are d ifferentiat ing. The e ndoder m is has been pns hed outward but not
dissolved though some oth er tissue has been dissolved. The group of cells in the cen ter are
from a ta ngent ial sect ion of a root. Note the activity in t he cambium a nd p ericyde and the
formation of new vascular eleme nts.
$ . Longitudinal section, the lower pa rt of which is about 1 mm . from the base of t he
cutt ing. The up vel' primord ium is just in side of t he pericycle. Lower primordium shows so me
di vision of t he pericy cle. At the base, ta ll gent ial J ivisions of the pericycle TOf m a rool. sti1\
lower .
C/. Longi lu dillal secl ion showing· a nxl!. differen tiated a nd a space d ue t o ~m rlial SolUl ion
of endodcfmi s.
C ONNA RD AND ZIMM ERM A N-PO R T U L ACA OLERACEA R OOTS
./
1
3
e
''l.'
;~I }
'. .._. k.;'-"
,;~
-
I,
' ,.Ii. .:'
i
/\ 1;
,1
JJ"
...
~'~.
<.
'I.
, .,
.; \ '
"
" .'
1
\ .
L
;,{ ' J~' .~ . . ,)
'; i J i
>
I. ,
· f
Ii
.
.f
. I.
~
:jl ' .
. .1
.
'..r._~; 1 '
',j f ' . .'
1
,
' .' 1
I
1 -.
, ,
\..,
P LAT E [ -
O IUGI N OF R OO TS
341
342
CONTRIBUTIONS FROM BOYCE THOMPSON IN STITUTE
[VOL·3
EXPLA NATIO N, PLATE II
All photomicrograph s are X47. Letters designa ti ng certai n structures have t he fol­
lowing meaning; e, endodennis; b, bast; p, pericycle; c, cambium.
1. Longitudinal section. (Most roots not m ed ia n .) Five roots , all at a bout same stage
developmen t, ha ve started one a bove t he oth er.
2. Longitnd inal section, showi ng a mature root which has dissolved a few la yers of cor­
tex and turned downward longitudin ally to emerge through cut surface.
3. Cross sect ion near base , showiJlg roots growin g tan gentia lly. Both a re probably in­
side 01 t he pcricycle.ln t he middle of t he cortex is th e cross section oC an older root wh ich had
originated higher up in the stem.
4. Cross section of a stem 0.7 8 mm . from the base, showing a root starting with in pith.
5. Cross section of a stern 0.98 mm . fro m t he base , showing the sa me b und le as 3. A
young root with a digestive pouch is differe ntiated inside of the bast fibers of the large bnn­
d ie which is just branch ing.
6. Cross section 0.88 m m . from th e base , showin g a root at right welt differentiated with
pouch. The eudodenni s is pu shed outward bu t not d issolved. The root a t the left is also in­
side of th e endodermis.
'93<)
CONNARD AND ZIMMERMAN-PORTULACA OLERACEA ROOTS
3
..I
..
5
PLATE II-ORIGIN OF ROOTS
343 344
CON TlUnUTI O NS FRO M BOYCE TH O MPS ON I NSTITUTE
[VOL ..1
Th e manner o f emergence of t he roots is peculiar . In most pla n ts ad ­
ventitious roots grow outward radially from the ce ntra l cylinder more o r
less perpe ndi cu lar to th e axis o f the stem and emerge th ro ug h the epi ­
d ermi s or pe rjde rm. In flu n;lane the roots emerge throug h t he c ut surface
and para llel to th e main a xjs of th e s t em (Fig. I) . Those nea rest t he base
aSS llme this longitudinal direc tio n from th e s t art. The maj ority of th e roots
start out rad ially as in o th er plants (PI. I, 3 a nd PI. II, r) but, when abo ut
ha lf th e cortex has bee n penetra ted, t he y turn down wa rd a nd proceed
lon gitudinall y (PI. II , 2 and 3) until they can ma ke their way th ro ugh th e
cut su rfa ce. \A/h e n one roo t is direct ly above an o ther, t he upper goes far­
ther o u t into the co rte:'\. a nd t hen tu rn s down wa rd. A crowdi ng of t he roo ts
may force the epider mi s to split and through th is opening root s sometimes
emerge. If enough of the stem cOrtex is rem oved by a tangent ial cut a ro ot
may come n ut nea rly pe rpendicu lar to the ax is . \Vhen, howeve r , the epi­
dermis with o nly a small portion of t he cortex is removed the roo ts a re no t
induced to continue their radial growth uut turn and go longitudin ally a s
desc ribed above (Fig. r, A).
T he facto rs con t rollin g thi s dir ec tion o f growth have not ye t bee n as ­
certa ined. The manner of e merg'ence cann ot be modified by changing the
position o f the cutting in re lation to grav ity. The experiments in which t he
epiderm is was removed as we ll a s t he observat ions on the na t ure o f th e
o uter tiss ues of t he stem see m to precl ud e t he possibility of ph ysica l ob­
str uction. Roots penetrate woody tissues or similar herbaceo us o nes in
many pla nts. Co lo rimetric deter mi nat ions of the pH values did not rev eal
a ny g radient from the interfascicu la r cam bium to th e epiderm is. There is,
possibly, som e chemical cond ition which favo rs g rowth in th e direc tion o f
th e cut s urfa ce . I n th is connec tion it may be no ted that in several in­
s tan ces when th e pith of th e cu tting was d ecayin g at the base, roots g rew
[award, o r act ua lly arose in , t he pith (PI. II , 4).
In th e normal purslane stem very littl e if any cell d ivi ~i o n takes p lace in
th e ca mbium. I n th e cut ting no callus is formed (PI. II , I a nd 2) a lthough
occa~i o nal cell di v ision s occur in the pith Or cortex. All the meri ~ tematic
activity o f the cutting, t herefo re, is associated with th e formation of ad­
ve nt itious roots. The hrst s igns o f meriste ma ti c ac ti vity occur in th e cam ­
bium two to four da ys after th e c utting is made. In yo unger stems t his
act ivity is usua ll y confin ed to cells of the interfascic ul ar cambium adj a ­
cent to th e bundles. The ce lls divi d e pe riclina ll y and a nticlinall y. In o lde r
ste ms t he activity is not loca lized at fi rst but d ivisions take place t hrou gh­
out most of the interfascic ul ar cambium (Plo 1,4) and o ften t he fasc icula r
ca mbium as we ll. In certain re gio ns, how ever, division is more rapid and
co ntinues {a nger so that soon de fi nite g roups of ce Us are differen tiated a s
in yo un ge r stems . O ften the ray cells between the ca mbium and pericyc le
may t ake part. In each case the primo rdia are in side t he pericycle, irregu­
193 11
CONNARD AN D ZIMMERMA N -PORTULA CA O L E RACEA R OOT S
345
ta r in s hape and size a nd s how no clear differentiation. The cells produced
by cambia l activity which d o not contribute directly to t hese new roots
may form the no rm a l sma ll connecting bundles or co nn ections linking the
vasc ular systems of root and stem (PI. I, 4). In well rooted cuttings the
rays become filled wi t h irregul ar masses of vascula r elements.
I n the next stage of root deve lopment the pericycle becomes involved in
t he activity (PI. I, 5). Th ere are a few tangential and a number of radial
divisions. These occur on ly above the primord ia a nd form the cortex and
epidermis of lhe roots. In P late I , 4, de fi nitel y differen t iated roots ma y
be seen ju s t wit hin t he endoderm is. Differentiation is not a lways evident
even at this period but becomes so a s the endodermis is pu shed outward
a nd dissolved (PI. I , 6). That solution rather th a n crushin g occurs seems
to be indicated by the fact t hat the growing roots do not appear to press
aga ins t t he tissue but rather to be separated from it by considerable space.
VARIATIONS FROM THE USUAL MODE OF DEVELOPMENT
There are numerous exceptions to the genera l mode of initia tion as out­
lin ed above. Not only is t he normal activity of the fasc ic ula r cambium
stim ula ted at tim es but root primordia occur within the bundle (PI. II , 5).
The exact tissues involved a re not known. S ince t he bund les ofte n branched
just a bove the poin t where s uch a primordium appea rs t he ray parenchy­
ma as well a s cam biu m or phloem ma y take part in forming these pri­
mordia . Since radia l growth of t hese primordia is restricted by the bast
libers, further growth mu st be in a longitudinal direction. In Plate II, 5,
a lt hough the bundle has branched , the bast still covers both parts. The
root in the narrow ray is well developed and the break betwee n it and the
fibers suggests that it has destroyed some of lhe cells within the bast fibers.
In the formation of s uch roots the pericycle cannot take part. Often also
the co ntribution of the per icycle is dis pensed with in roots w hich from the
earliest divisions grow tangentially rather than radia ll y. In t hese cases the
pericycle appea rs t o have been dissolved a nd even wit h roots normall y
oriented , the space betwee n t he root and the stem whil e t he endoderm is is
still un to uched may mea n solu tion of at least p art of the peri cycle (PI.
11 ,6). The root would the n be formed entirely from camb ia l de rivatives or
t he peri cycle would con tribu te only part of the cells. In a ll t he examples
just mentioned the peri cycle does not necessarily make the outer tissues
of t he root. Likewise there is n o evidence that ro ots origin ate from the peri­
cycle alo ne. As men tioned a bove, roots in two cases were fo und to have
been fanned entirely within t he pith (PI. fl, 4) a nd un conn ected w ith the
cam bium. Wi th this exception t he activ ity of the interfascicular ca mbium
appea red to be th e most essen t ia l feat ure of the formation of t he ad venti­
tious roots of purslane.
CO NTRIB UTIONS FROM BOYCE THOMPSON IN STITUTE
[VOL. 3
SUMMARY
1. The maj ority of adventitio us roots from c uttings of Portulaca
oleracea L. arise in the medullary rays within 5 mm. of th e base, and emerge
through the cut s urface without penetrating the epidermis.
2 . The earliest meristematic activitY' is in the interfascicular ca mbium.
Cells derived from t hese div isions form the inner por tions of the root and
accesso ry vascular tiss ues of the stern .
3. The pericycle generall y forms the outer portions of the root , the e n­
dodermis being the hrst t issue dissolved by the root.
4. In stan ces occur in which t he roots a re fully differentiated a nd some
tiss ues a lready dissolved with the endodermis still intact. Some primordia
·were found within vascular bundles where the pericycle could not take
part in their forma tion.
5. U nder some conditions roots a rise within t he pith.
LITERATURE CITED
J. H.,
a nd SWINGLE, CHARLES F. Vegetative propagatjon from t]le stand­
poin t of plant a natomy . U. S . Dept. Agric. Tech. Bull. lSI. 98 pp. 1929.
2. TIE GHE::'f, PH. VA N, et DOULIOT, H. Recherches comparatives sur I'origine des meJO bres
endogf;ues dan s les pJa ntes vasculaires. Ann. Sci. Nat . Dot. Ser. 7. 8: 1- 660. 1888.
1. PJUESTLEY,