r9751 TOMLINSONAND SODERHOLM:CORYPHA The Flowering and Fruiting of Corypha elatain South Florida

P. B. Tolrr,rNsoNl AND PAUL K. Soonnuor,m2

n.: The flowering of a specimen o{ C. elata and C. umbraculilera L. which Corypha elata Roxb., P.I. 74425, in the established use{ul diagnostic features. summer oI 1972 at the Subtropical Hor- There appears to be less information ticultural ResearchStation, U.S. Depart- concerning other species. ment of Agriculture (formerly United StatesPlant Introduction Station, Chap- The Specimen man Field), Old Cutler Road, Miami, The palm in question (Fig. 1) origi- Florida, the first known {lowering in the nated from a batch of seeds{rom the continental United States,permitted the Philippine Islands received following observationson the flowering at Chapman Field process and a detailed analysis of in- on luly I7, 1927, which germi- florescence construction in the genus nated shortly thereafter, were moved to Corypha. Some account of the bioener- eight-inch pots in 1932 and getics of the palm is also made which planted out in 1935. At the time of allows us to assessthe extent to which flowering in June 1971 the palm was 44 biomass is diverted into vegetative and reproductive processes. Although the Cory.pha inflorescence has been mentipned frequently in bo- tanical texts as a dramatic example of {lowering becauseof its overall size (see also Hodge, 196I), the branching pat- tern has never been describedin detail. The presentarticle in part doesthis and affords a comparison with inflorescence construction in two other hapaxanthic (once-{lowering) palms, Nannothops (Tomlinson and Moore. 1968) and Metroxylon (Tomlinson, l97l). The overall appearanceof the inflorescence of Corypha is indicated by Blatter 11926) and elaboratedby Douglas and Bimontoro (1956) who madea point-by- point comparison of inflorescencesin

'Harvard University, Harvard Forest, Peter- sham, MA 01366. " U.S.D.A. Subtropical Horticuitural Research Station, 13601 Old Cutlcr Road, Miami, FL 1. Overall view of specimen in flower, July 33I58. 27, I97I. lVol. 19 B4 PRINCIPES (April 1973). Table1. corypha elatain flower,46 yearsold at time ol measurentent

Overall height 19.3 m \63' 6") (48'10") Height to lowest inflorescence branch I4.9 m Height of inflorescence 4.5 m (14' 8") Diameter breast heighr (3' 6") 0.86m ( 2'I0") Diameter immediately below inflorescence 0.46m ( 1' 6") Total volume of palm (calculated) ,.5.1m3 Total number of scars visible on above-ground c. 350 part of trunk

years old from seed. It is not known if order branch (9axr-see below) was other palms from the same seed source cut to obtain a measurementof {ruit o{ are in the Miami area. As indicated weight. Unfortunately' no samples below, the period of flower presentation {lowers were obtained; the illustrations (Fig. is quite short but the fruiting processis o{ which are included 9E, sp' extended; seed continued to ripen for F) are from a collection oI Corypha E. about IB months' In March 1973 all (Read s.n.) supPlied via Dr. H. had {allen and the palm was {elled Moore, Jr. Table I summarizesrelevant so that the inflorescencecould be ex- dimensions. amined before it decomposed. Previ- Foliage leaveshave an encircling leaf in the manner ously, in February L972, a single {irst- base,split down the back

d'euelop' Ta,ble2. Corypha elata. Sequenceol euentsin later stagesol inllorescence ment, 1971.

which is still intact. June l0 Main axis (axs) extended well above crown First-order branches (ax1's) well extended' Few second-order branches (ax:'s) exposed (Fig. 2). second- and June 16 First-order branches completely extended, numerous third-order branches exposed; rachillae evident (Fig' 3)' (Fig' 4)' June 18 Very first flowers at anthesis; crown still intact June 2I Numerous {lowers at anthesis. to droop' Itne 22 Lower pendulous, upper leaves beginning flowers at anthesis. Irne 29 In{lorescence in full {lower; maximum number of July 2 Most leaves drooping. some rachillae ap- July 18 Many flowers fallen, but flowering continuesl pearing bare. July 25 Flowers still at anthesis; all leaves drooping. JuJy 27 Young fruits first evident (Fig. 5). August 20 Crown completely collapsed, all leaves pendulous; young fruits conspicuous (Fig. 6). September15 In{lorescence densely covered with fruits of almost {ull size. Leaves shrivelled, no longer green' r9751 TOMLINSON AND SODERHOLM: CORYPHA B5

2. June 1lth: {irst-order branches well extended and completely enclosed by tubular bracrs,

June 17th; extension of first-order branches complete, second-order axes extended to expose the rachillae. tVol. 19 B6 PRINCIPES

before beginning of flower opening; crown June ISth; overall view of inflorescence shortly still intact and leaves erect.

of a number of fan palms (-type leaf base o{ Tomlinson, 1962)' The petiole is grooved adaxially and has stout, coarse, marginal teeth, in contrast to the finer teeth ol C. umbraculilera (Douglas and Bimontoro, 1956). The blade is costapalmate. Flowering The serieso{ photographs (Figs. 1-6) illustrate the later stagesof in{lorescence developmentwhich began in early 1971 but for which we have no precisedate. The time sequenceof signi{icant events is provided by Table 2. These observationsshow that the pe' riod of actual anthesis (pollen presenta- tion) for the whole inflorescenceis quite short in comparison to the palm's total life. Beginning on June l8th, all flowers seemto have completedtheir expansion and pollen sheddingby the end o{ Julv. Corypha elata is clearly self-compatible 27th; late stage, most flowers fallen; 5. July was the only palm in flower crown collapsed, leaves pendulous. since this 19751 TOMLINSON AND SODBRHOLM: CORYPHA 87

6, August 20th; young fruits approachingfull size, in South Florida and viable 'We seed was tailed observation to support this sug- set in abundance. have no precise gestion. The proportion of viable seeds knowledge of the pollinators; bees vis- in the crop seemsvery high and prob- ited the flowers in considerable num- ably approached 100 percenii Details "of- bers, but the odor is described as of total productivity are given later. An fensive" by Douglas and Bimontoro. illustration of the amount of on one branch o{ the inflorescenceis eiven Fruiting in Figure 7. Fruit enlargementwas rapid, but the overall fruiting period was quite lengthy. InflorescenceConstruction Although fruits of apparently {ull size The Corypha inflorescence has the were presentfrom mid-September(three samesimple principle of construction as months after the peak period of anthe- has been described fior Nannorrhops sis) at which time the inflorescence (Tomlinson and Moore, \968) and Me- branches were densely crowded with troxylon (Tomlinson, L97I) . The vege- green fruits, these remained on the tree tative axis is continuous with the main {or another year before they }regan to axis of the inflorescenceand there is a fall, suggestinga long period of embryo transition from foliage leaves through maturation. We have no further de- leaves with reduced petiole and blade PRINCIPBS

"crop" is in the t'wo galvanized Febrnary 9,1972; {ruiting branch (9 axr) cut from the palm' The bucl

(ax") infloresccnce, 8. Corypha elata, details of inJlorescence construction. A. Main axis oI entire cnd of this *ith uiifirsr-order branches cut oJf near their insertiorr (a iz2aapprox')' The clistal 19751 TOMLINSON AND SODERHOLM: CORYPHA 89

D

qx1

rochillo

oXz

is drawn in detail as Fig. 94. B. Entire first-order branch (13 axr) with all large second-order branches except one removed (x ltd. C. Details (x 1/d oI insertion oI a fiist-order (ax1) (axo), branch on the main axis subtending bract (brr) removed (its scar represented bv the stippled band). This shows the overlapping prophyll (1 br,) and second brao (2 br,), boih of 1rO which are empty. D. Details (X of insertion of a second-order (axr) branch on a first-order (ax'), axis subtending bract (br:) split down back; the base o{ the branch shows no modified 1/+) prophyll. E. Details (X of a third-order (ax") branch inserted on a second-order branch (ax"). There are {ourth- but no fifth-order branches on this system. All axes end in flower-bearing (X 1/il rachillae. F. Details of a distally inserted firsr-order branch (3g axr) in which thi prophyll (l br,) subtends a second-order branch (ax,)-cf. Fig. gC. ls 90 PRINCIPES [Vol. main axis of to modified leaves or bracts (Fig. BA). The lowest leaves on the a dorsal cleft Any of these organs, whether modified the inflorescence have branch or not, that subtends a first-order branch through which the subtended is contrary of the inflorescence is here considered protrudes. This observation Bimontoro a bract for purposes o{ numbering. to that of Douglas and that in Corvpha Bract is also an appropriate term to (1956) who emphasized protrude re{er to modified leaves on second and elata the first-order branches sheath,in con- subsequent orders of branching. through the moulh of the where It is convenient to refer to the parts o{ trast to Corypha umbraculilero the dorsal cle{t. this system o{ branches by symbols; ax they prbtrude through a good diagnostic {or axis or branch, br {or bract' Parts This is probably not the two may be indicated by numbers, e'g. axg- feature by which to distinguish can occur main axis, €Ix1. 3X2. ilX: . . etc. {or speciessince both conditions on the first-order first, second, third etc. order of on one axis. Bracts (brz's) tubular and papery, branch. Bracts may similarly be desig- branches are bracts (br1's) on nated according to the order of branch and resemblethe distal which they subtend, e.g. br1 is a bract the main axis. give impressionof the on the main axis (axe) subtending an To an accurate inflorescence,it is repre' axr, brz would subtend an ax2 etc. skeletonof the to scale in Fig- Any axis can then be identi{ied by num- sented diagrammatically the branchesare drawn bering from the base, e.g. 9ax1, the ninth ure 10. Here all plane, in {act, they are first-order branch. in one whereas, which corresponds Internode length along the main axis arrangedin a spiral phyllotaxis. The most changes little in the transition from the to that of the 3/B of this spiral are certain vegetative to the reproductive phase, ex- evident features parastichies. One serieso{ these can be cept that distal internodes are shortest. in Figure BA, which repre- There is. however, a rapid decrease in seenclearly -'successive the main axis with each iliameter of internodes. In sents most of cut off a little above Figure 10 internodes have all been first-order branch These parastichies probably drawn of equal length. There is a rapid its base. those which are evident increase in rate of extension relative to correspond to trunk and which serve the vegetative phase so that in a matter on the vegetative feature for" this species of a few months the panicle stands above as a diagnostic Bimontoro, 1956)' the old leafy crown. The reduction in (Douglas and bract subtends a single leaf size previously mentioned, resulting Almost every are empty bracts in tubular, sheathing bracts, is clearly branch; the exceptions axis (Fig. correlated with decrease in internode at the baseof each first-order proximal diameter. BB. C). On the first-order

)

(x of main g. corypha elata, further details of inflorescence constluction. A. Termination %) in dotted out' uxis (a*"); its continuation into a rachilla is shown (arrow). Branches are shown 'iermination (13 in Fig. BB; its line. B. (x )/d of the first-order branch ax') illustrated C' Single continuation into a rachilla it .ho*tt (arrow). Branches are shown in dotted outline' D' Portion of a flolver cluster (X B) \rith flowers represented by their scars; bracteoles evident' sp. (x 9/2), rachilla (x 2) with spirally-arrang"d flo*"t "irr.t"r*. E-G. Flowers of corypha (G). flowers are from from above (E), in longitudinal seciion (F), and as a floral diagram These to flower cluster on , ,s"p"r.r" colie"tion, F-ead, ,.r. Kingsron, Jamaica. H. Ripe fruit attached a rachilla (X I). r9751 TOMLINSON AND SODERHOLM: CORYPHA 9I

J! ;1

^.1.r..1',tlt':;'

--->

qx1- 19 92 PRINCIPES lVol. them- branches the first two bracts are com' seen in the first-order branches (38 monly empty (Fig. f l). The first selves. A distal first-order branch (1 brz) is a distinct bicarinate prophyll axr) is shown in Figure BF; here the prophyll a branch and the tFig. BC). Second-orderbranches (axz) subtends lack such a modiiied prophvll (Fig. bracts are scarcely tubular. Details of (axo) is 8D). There is an abrupt transition in the end of the main axis itsel{ of bract shape from first- to second-order shown in Figure 9A', and of the end branches. On the first-order branches, a first-order branch (13 axr) is shown to bracts are tubular, except {or distal in Figure 98 to-indicate the degree ones; on second-order branches, the which massiveaxes become reduced dis' bract is representedmerely by a short tally. There is usually a single empty scaleor even a ridge of tissue (Fig. BD, bract below the ultimate branch system E). An entire first-order branch with (x in Fig. 9A, B). Each axis terminates all but on€ ax2cut off is shown in Figure in a rachilla, following the rule we have is BB; similarly, a single first-order branch pointed out; this terminal rachilla (arrows) is shown to scalein Figure 11, but with shown in Figure 9.A and 9B to the second-order branches along it with the branches in dotted outline shown in one plane. contrast with the parent axis' It is re- Branching continues to the fourth markable that tle apical meristem ot order on most second'order branches the palm, which in the vegetative con- to three feet (Fig. BE) ; branchins to the fifth order dition generatesan axis up is uncommon and occurs only on the in diameter, finally terminates in the by lower parts of second-order branches, narrow flower-bearing axis indicated as shown in Figure 12' To indicate the arrow in Figure 9A' further the constructional principles, a The flowers are borne in clusters first-order branch (9axr) is drawn out whose detailed morphology represent pers. to scale in Figure' 11 while a second' modified cincinni (Uhl, comm.) of order branch-'is shown diagrammatically the type which occurs in other coryphoid in Figure 12. Comparison of the dia- palms (e.g. Uhl, 1969). The significant grammatic figures (Figs. 10-12) with functional unit may be the flower cluster the actual representation of parts in itself as much as the individual flower, to a Figures 8 and 9 should allow one to and we largely restrict ourselves this unit in subsequent reconstruct mentally the overall archi- considerationo{ Flower clusters are elongated tecture of this in{lorescence. analysis. and arranged in regular spirals along the rachilla (Fig. 9C, D). Bracteoles Rqchillqe are associated with the flower clusters, A11axes end in rachillae, i'e. the ulti- on the basis o{ one bracteole to one mate visible flower-bearing axes, as de- flower; the whole cluster itself is sub- {ined by Tomlinsonand Moore (1968). tended by a single bract, representedby They are surprisingly uniform and aver- little more than a crescentic ridge of age 25 cm. in length, which is help{ul tissue. Figure 9C represents an old in subsequentcalculation (Fig. BE). flower cluster from which all {lowers Distal parts of both the main axis have fallen, only their scars remaining and all first-order branches are much without setting fruits. There seem to be modi{ied and resemble a second-order from 5-10 flowers per cluster, but the branch system,as is evident in Figures number of {unctional flowers is difficult 94" B. Similar modifications mav be to establish. r97sl TOMLINSON AND SODERHOLM: CORYPHA 93

40

F10 F

10. Corypha inflorescence in toto. The main axis and first-order branches (axr) drawn to scale. The branches are all represented in one plane (not in the shallow spiral on which they are actually inserted-cf. Fig. BA). Dotted lines represent incomplete branches or those whi"li -e." ,rot measured. Numbers to far left and right number axes successively,counting lowest as l. Numbers near the insertion of each branch represent the total number of branched, second-order branches (axr) on each {irst-order branch (ax'), i.e. corresponding to the one drawn into Fig. BB.

Flowers qnd Fruits diameter at maturity with an average dry weight of 2.3 gm. In Flowers are perfect, with the parts in old rachillae the sites of ripened but 3's but uith 6 slamens. An example is fallen fruits are easily identified by the persistent, shown from above, in vertical section en- larged and flattened calyx. and as a floral diagram in Figure 9E, These fruit scars are used in the presented F, and G respectively. The three carpels counts subsequently. are fused; each contains a single ovule but normally only one of the three func- tions. Usually only one fruit per flower QUANTITATIVEASPECTS cluster matures (Fig. 9H) ; rarely are So far we have established the con- there two. Fruits are about 1.5 cm. in structional principles of the Corypha tVol. 19 94 PRINCIPES

all b-ranchedsecond'order lI. One first-order branch (9 ax1) dlawn to scale, showing the length oI they actually describe. frur"h". (axr). Branches are shown in one plane, not in the irregular spiral branches (ax") on Numbers towards the end of each branch indicate the number of third-order the length (in cm') oi "u"tr-*"a-otder branch. Numbers at the base of each branch indicate BB' the subtending bract. This branch corresponds to the entire unit illustrated in Fig'

inflorescenceon a- qualitative basis. A Total number of {irst- order branches : 48 series of measurementswere made of !features Total length of all quantitative from representa- {irst-order tive sampleswhich permit us, by a series branches = 97.65m. (320.3 {t o{ appioximations,to arrive at the over- Averagenumber of all reproductive potential of this in- second-order per The calculations are a little branches florescence. first-order involved, and only the results, with some branch = 13.3 indication of how they were arrived at, Total number o{ are included here. second-order branches : 640 Average number o{ Flower Number rachillae/second- order branch = 33 is obtained {rom in- Flower number Total number of {ormation about the total length o{ rachillae - 21,720 flower-bearingaxes on the inflorescence; Average length of one rachilla = 25 cm. effectively an estimate o{ the total length pro- Total length of o{ all rachillae was rnade. This is all rachillae - 5,280 m' vided by the {ollowing in{ormation, Average number of addedto an estimateof the total number flower clusters/ rachilla = 75 and second-order branches of first- Estimated total (omitting branches 49-SI) taken from number of flower Figure 10, as follows: clusters : 1.58X10u t9751 TOMLINSON AND SODERHOLM: CORYPHA 95

12. Corypha inflorescence, ultimate branch complex, an actual example, but not dPdwn to scale. This corresponds to the single branch unit drawn in detail in the middle of Fie. 8B. Ultimate units (rachillae) are represented diagrammatically.

Estimatedrange of 9C suggests seven flowers on the basis numbersof pos- of their scars). We have seenthat each siblefunctional flowers = 3-l5xl06 flower cluster normally produces, at most, one fruit, so that it may be hy- The total number of flowers is not pothesized that the successful fertiliza- easily obtained, regardless of how pre- tion of one flower in a cluster causes cisely one is able to estimate total num- the inhibition of further maturation of ber of flower clusters. becausethe num- all younger flowers. In the above calcu- ber of actual flowers in each cluster lations we have used a range of 2-10 varies, but more importantly because a flowers per cluster in arriving at a knowledge of the number of flowers in range of final figures. each flower cluster which reach func- On this hypothesis the total number tional maturity is quite imprecise (Fig. of flowers reaching anthesis per cluster tVol. 19 96 PRINCIPES of is simply a function o{ the age before fruit (2.3 gm.) into our estimate successful pollination of any flower total fruit weight. The values arrived within each cluster. This is why it is at are 242,800 and 355,290 fruits re- biologically most meaningful to con- spectively. These figures are su{ficiently sider the {lower cluster as the significant closeto suggestthat our approximations reproductive unit. representsomething rnore than well-edu- Estimating flower number on this cated guessesand indicate that the in- quarter basis we have arrived at a figure o{ the florescence prodpced about a order of t0 X 106 for the total number o{ a million {ruits. The values arrived of flowers per inflorescence (i.". l0 at for {ruit weights are: a) from drY ks'; million). This figure, having been ar- weight of a large sample:8L7.2 rived at by a series of approximations, b) from an estimated 15 Percent of = can give us no more than an indication flower clusters which form fruits of the order of magnitudein the number 545.1 kg. o{ {lowers produced, but the very large number involved can be grasPed. Of Bioenergetics particular interest is the total length of We have so {ar establishedreasonably flower'bearing axes (i.e. rachillae) on accuratefigures for the total amount of "energy") this inflorescence; our {igure is 5,280 dry matter (or which goes m., i.e. if all the rachillae were laid end into the production o{ {ruits in the flow- to end they would stretch more than ering of this hapaxanthicpalm and have three miles!!! produced a minimum value of 545'1 kg. "What A more precisefigure is obtained for It is interesting to ask propor- the total dry weight of the fruits in this tion of the total biomass produced by in{lorescence,since a whole first-order the palm in its lifetime goes into its branch (9 axr) with fully enlargedfruit reproductive ef{ort?" A series of mea- was cut from the palm on Febrtary 1972 surements as follows gives us some indi- (Fig. 7). The total weight (wet) of cation of this value: fruits from this branch was measured Total trunk biomass = 863.5kg. togetherwith the dry weight of a smaller Total leaf biomass = 2,023 kg. sample. An estimate of the fraction of Total root biomass total inflorescencerepresented by this the (estimated) 288.6kg. axis is easily obtained from the in{orma- - Total inflorescence tion in Figure l0 and, assuming a con- biomass = 49.3 kg. tinuous proportion between axis length Total biomassof fruits = 545.1ke. and numbers of fruit borne, allows a calculationof total fruit weight (wet and Total biomass :3,769'5ks. dtv). The value for the total fruit These values are obtained crudely, as weight (dry) obtained in this wav is follows: 8L7.2 ks. We have calculated the total number Trunk bioma.ss.From an estimateof (calcu- of fruits produced by two di{{erent the total volume of the trunk and treating means. On the one hand, by obtaining lated from its dimensions regular truncated cone) and an averageof the number of fruit scars it as a a knowledge of the dry weight o{ a per rachilla ( 1I.5 ) , and multiplving small sample o{ calculated volume. this value by the total estimatednumber of rachillae and, on the other hand, by Total leal biomass. From an estimate dividing the averageweight of a single o{ the approximate total number of r9751 TOMLINSON AND SODERHOLM: CORYPHA 97

leaves (counts of visible leaf scars. From these calculations. we can see plus estimated 10 percent for leaf that, if our measurements are at all re- biomass produced in juvenile stages, liable, the Corypha palm expends about i.e. before a visible trunk is devel- 15 percent of its total energy budget on oped) and a knowledge of the dry reproduction. The value should be cor- weight of a leaf taken from a culti- rected by a factor which accounts for vated specimen. loss of energy via respiration during the vegetativg period of growth, i.e. the Total root biomass. A rough estimate total biomass is underestimated by this of 10 percent of total leaf and trunk factor but the value (e.g. l0 percent) biomass. would reduce the final percentageby a Total inllorescence bioma.ss. From a small figure. knowledge of the dry weight of sam- ples of the inflorescence and the in- coNclustoNs formation contained in Figure 10. From the above calculations. it can No account is taken of the weight of be concluded that about 15-20 percent flowers, which were not available in of the total dry matter produced by the our sample. Corypha palm during its lifespan of It is well known that starch accumu- about 50 years is eventually directed Iates in the trunk during the vegetative into its reproductive effort. These fig- period and is rapidly mobilized and ures may seem inordinately high and utilized during the flowering process, could suggestthat our measurementsand offering an interesting subject for study the calculationsbased on them are sus- by translocation physiologists (van Die pect, which is possible in view of the and Tammes,1975) . Corypha itself has small samples obtained and the approxi- been used extensively as a source of mations made. However, some corrob- sugar and gventually alcohol by artificial orative evidence is available. Burkill rapping (1935) recordssome 200 pounds (about 'We lGibbs, 1911). can divide this mobilized reserve 90 kg.) of starch in a palm which may into that part which goes into the con- be extracted as sago. This isolated fig- struction of the skeleton of the in- ure seems too low to account {or the florescence and that part which goes dry matter which is finally incorporated into the fruits. On this basis we have in fruits but since we dd not know the the following figures: size of the palm measured,it is not very helpful. Total biomass of the The productivity of Corypha is more Coryphapalm = 3,769.5 ks. helpfully indicated by Gibbs (1911) Total biomass of fruits = 545.1 kg. who measured.2,699liters o{ sap tapped Total biomass of from a tree with a volume of 2.71 cubic inflorescence = 49.3 ks. meters in a period oI L32 days. The sap Total reproductive contained 252 ks. o{ sucrose. Average biomass = 594.4 ks. yields for other trees ranged from 180- Percentage of total 324 ke. These values are somewhat I. Dtomass rn replo- lower than our calculated figures but it ductive biomass = 15.77% must be remembered that they are based [If the higher value of fruit biomass on trees manipulated artificially in order (817.2 kg.) is used, the result is to induce mobilization of reserves;this abofi 22%l mobilization is likely to be less efficient 19 98 PRINCIPES tVol. than during the natural processo{ fruit (of the order of 50 percent) on the an' development. Also the tree measured nual basis may be measured,i'e. in any in detail was much smaller than the one one year the amounts of drY matter we have measured (2.71 cubic meters partitioned between fruit bunch, on the ,l compared with 5.1 cubic meters). Our one hand, leaf and trunk, on the other, results thus seemto be of the right order are about the same. On this scale, I of magnitude. Corypha is seen to be less efficient than Furthermore, when other are at least one ot[e.r palm. The figures considered,the valueswe have produced quoted fior Ewterpe globosa by Bannister seemto represent gratifyingly close esti- (I970)'are less completebut we have per- mates. Long-lived woody plants which usedthem to arrive at a value of 4.5 are monocarpic (once-flowering) are cent for the reproductive biomass in uncommon. The condition is familiar this species. in a number of (e.g. Seedpredation is known to be a sig- somepalms, Agaae), a few dicotyledons nificant factor in survival of plants with (Hall6 and Oldeman, 1970, p' 20), and large seeds(Harper et a1.,1970). There has recently been described by Veillon was some evidence of beetle attack on (I97I) for a branched apocynaceous fallen (and even attached) {ruits of rree, Cerberiopsis candelabrum Yieil' Corypha, but the high viabilitv of seeds lard, native to New Caledonia. Other- and the enormous numbers Produced wise. one has to turn to annual weeds would seem to ensure survival of the for a comparable reproductive strategy. palm. Production of large numbers of Here, since the continued survival of the seeds at wide intervals is probably an species is entirely dependent on seeds, e{fective way of combating seed preda- with the vegetative lile span of each tion. However, measurementsneed to individual very short, a large percentage be made on natural populationsof Cory- of the total biomassof each is ex- pha with an understanding of the ecol- oore- pended on feproduction, i.e. the ogy of palms in wild sites. productive etIorr" of the plant is high. For the present, our very tentative For annual, weedy species,values of the measurements suggest that despite its order of 14-35 percent seem typical massive size, Corypha has to function (Harper and Ogden, 1970). In Peren- .essentially as an annual weed in order :)' nial plants, e.g. trees, the values are to survive. much smaller than these although there seem to be no accurate measurements SUMMARY 1970). In its re' (Harper and Osden, A Corypha elnta Pilm which flowered is productive effort therefore, Corypha and fruited during I97l-72 alter 44 "annual" is comparable to an weed, but years of vegetative growth was estimated suc- unusual for the time span between to have producedon the order of l0 mil- (more years) cessiveflowerings than 40 lion flowers on flower-bearing axes of a plant and the stature that the achieves total length of 5,280 m. (more than three (about 20 m.). miles). The total number of ripe fruits It is worth making a more direct com- was about a quarter of a million, with parison with other palms. Precise fig- a total estimated dry weight o{ between ures for the African oil palm (Elaeis 545 and Bl7kg. (about1,200-1,800Ibs.) guineensis Jacq.) have been provided which represented between L5 and' 22 by Corley et aI. (L97L), Hatdon et ol. percent of the total dry weight produced (1972) which show that higher values by the palm during its total life span. r9751 TOMLINSONAND SODERHOLM:CORYPHA 99

The palm is comparable to an annual flowered in the Hortus Bogoriensisduring weed in its division of biomass between 1953to 1955. Ann. Bogor. 2: 137-148. Grnns,H. reproductive and vegetative phases and D. 191I. The alcohol industry of the Philippine Islands. Philipp. J. Sci. its massive fruiting over wide intervals 4.6:99*206. of time is probably a very effective way H.rnf, F. .lNn R. A. A. Ororlrlx. LW\. of surmounting seed predation. Essai sur I'architecture et Ia dynamique de croissancedes arbres tropicaux. Mas- son et Cie. Paris. Acknowledgements Hennor, J. J., R. [k {. Conlny, lxn S. C. We are indebted to Priscilla Fawcett" Oot. {972. Analysis of growth in oil palm. II. Estimation of genetic Botanical Illustrator, Fairchild Tropical variances of growth parameters and yield of fruit Garden for Figures 8 and 9, to Dr. bunches. Euphytica 2l: 257-264. Martin H. Zimmermann. Harvard Forest Henrrn, J. L., P. H. Lovnlr,, .lxo K. G. for Figure 7, and to Dr. Dent Smith for Moonr. 1970. The shapes and sizes of the photograph which forms the cover seeds. Ann. Rev. Ecol. Syst. 1: 327-356. Hennrn, L. eNn OcrnN. 1970, picture. J. J. The re- The assistanceand interest o{ productive strategy of higher plants. I. Mr. W. E. Manis, previously Horticul- The concept of strategy with special refer- turist-in-Charge, U.S.D.A. Subtropical ence to Senecio aulgaris L. J. Ecol. 58: Horticultural Research Station is appre- 68I-698. Hoocr, provided W. H. 1961. Nature's biggest bou- ciated. Field assistance was quet. Principes5: 125-134. by Gary Vitale and George Wilder. TourrNsox, P. B. 1962. The leaf base in palms, its morphology and mechanical LrrunerunnCrrso biology. J. Arnold Arboretum 432 23-4,6, 1971. Flowering in Metroxylon (the B-r.unrsrnn,B. A. IW|, Ecologicallife ,cycle sago palm). Principes 15: 4942. globosa of Euterpe Gaertn. pp. 8299-314. Torur,tNsolr,P. B. lrvo H. E. Moonn,Jn. 1968. In: Odum, H. T. and R. F. Pigeon, A Inflorescence in Nannorrhops Ritchiana tropical rain forest. Division of Technical (Palmae). J. Arnold Arboretum 49: 16- Information, U. S. Aiomic Energy Com- 34. ,rnission. .t Urrr,, N. W. 1969. Anatomy and ontogenyof Br.rrrrn, E. 1926. The palms of British the cincinni and flowers in Nannorrhops and Ceylon. Oxford University Ritchiana (Palmae). J. Arnold Arboretum Press. London. 50: 411-431. Bunrrr-r,, I. H. 1935. A dictionary of the VeN Dro, J. emo P. M. L. Teiranars. 1975. economic products of the Malay Penin- Phloem exudation from monocotyledonous sula. 2 Vols. Crown Agents. London. axes. fz: Encyclopedia of Plent Physiol- Conlrv, R. H. V., J. J. H.rnoorv, ar,lo C. Y. ogy, n.s. vol. 1, Zimmerman, M. H. & T.s.tv. 197I. Analysis of growth of the oil Milburn, J. A. (eds.). Springer Verlag, palm Elaei.sguineensis Jacq. I. Estimating Berlin (in press). o{ growth parameters and application in Vnrr.r.oN, J. M. 1971, Une Apocynac6e breeding. Euphytica 20: 304-315. monocarpique de Nouvelle-Calidonie Cer- Doucr,es, J. lrvn R. R. BrMoNrono. 1956. beriopsis canilelabrum Vieill. Adansonia, Identification of the Corypha palms which series 2" 1I: 625-639.