DAY-LENGTH AS RELATED TO VEGETATIVE DEVELOPMENT IN ULEX EUROPAEUS L.

DAY-LENGTH AS RELATED TO VEGETATIVE DEVELOPMENT IN ULEX EUROPAEUS L. II. ECOTYPIC VARIATION WITH LATITUDE BY L. H. MILLENER University f Auckland, New Zealand {Received 27 September 1961) (With 2 figures in the text) SUMMARY The extent f the juvenile phase in grse seedlings grwn under unifrm cnditins frm seed cllected frm varius statins thrughut Britain, has been shwn t vary significantly with the latitude f the parent plants. Regular seasnal variatin in the juvenile-adult balance f seedlings has als been demnstrated. Day-length seems t be the nly envirnmental factr which can be pstulated t explain such variatins fr it alne changes cnsistently with latitude and frm seasn t seasn. INTRODUCTION Studies in the prtean species Ulex eurpaeus (cmmn grse r furze) extending ver several years, have shwn almst bewildering variatins in grwth-frm and in branch, spine, leaf, flwer, fruit, seed, embry and abve all, seedling characters. As part f mre general researches n the British species f Ulex carried ut at the Btany Schl, Cambridge, seedling ppulatins f grse were investigated in Britain n a reginal basis. It was knwn that in New Zealand, and f curse elsewhere, climatic-latitudinal changes may have the mst prfund influences upn the habit and mrphlgy f certain plants (e.g. the mangrve, Avicennia resinifera L.). MATERIAL AND METHODS Sixteen difl"erent seedling ppulatins, representing fifteen lcalities frm almst the whle length f Britain, were grwn in bxes in a mdified Jhn Innes cmpst. The seed was swn n 20 June and by the end f September all seedlings had cmpleted their juvenile develpment. The cmpst (Lawrence and Newell, 1942, p. 59) was made withut lime and with the peat fractin dubled (half being Irish Sphagnum peat, half a black, acid peat frm Essex): its ph was 5.4. It was inculated with grse rt-ndule bacteria as described in an earlier paper (Millener, 1961). The psitins f the seedling rws in the bxes were permuted s as t avid 'end effects'. The bxes were interchanged in psitin accrding t a plan which ensured the neutralizatin f the efl:ects f any unevenness in the distributin f light, temperature, etc. It was nt practicable t grw ppulatin replicates in this experiment: but, apart frm the statistical checks made, many ther experiments invlving replicatin have demnstrated that the results adduced here can be relied upn. 119

120 L. H. MiLLENER A wnderful range f variatins, invlving all vegetative features f the structurally cmplex grse seedling was exhibited: but attentin will be fcused here n the 'transitin numbers'; that is, the mean numbers f juvenile (essentially cmpund) leaves per seedling (Table i). The transitin number, as has been shwn elsewhere (Millener, 1961), is the best measure f the extent f juvenile develpment in a ppulatin. Transitin numbers can nly be usefully cmpared when ppulatins are grwn at the ne time and in the ne medium. The 'Ordinary' Hindhead and the Breckland ppulatins (Table i), fr instance, prduced a mean f 14.3 and 14.0 juvenile leaves per plant respectively: in an earher experiment, when the seedlings were ver-wintered in the juvenile cnditin, and then allwed t grw n, the means were, respectively, 40.2 and 22.9 (Millener, 1952). Differential seedling respnses such as these have ften heen bserved in experiments n Ulex. The results f early wrkers (e.g. Wager, 1897) ften cannt be evaluated because factrs such as this were neglected. RESULTS As will be seen frm Table i, the varius grse lcalities may readily be put in five grups n a latitudinal basis. If the transitin numbers in the last clumn be examined, a striking fact emerges at nce. Althugh there is cnsiderable variatin amng means within certain f the latitudinal grupings, and sme verlapping f values frm grup t grup, there is a steady rise in the pled means frm the mre sutherly grups t the mre nrtherly. Even thugh the means are nt all significantly different frm ne anther the seriatin is mst impressive. Whatever the statistical significances, the randm chance f finding the means as they are wuld be i : 120 (Factrial 5) against. The imprbability that the latitudinal-mrphlgical crrelatin is due merely t chance is heightened if the transitin numbers f New Zealand and Hildersham (Cambridgeshire) grse seedling ppulatins, grwn tgether between May and July in essentially the same medium as is described abve, be cnsidered. The 400 N.Z. seedlings (frm parents grwing in Lat. 37 S.) develped nly 12.4^0.17 juvenile leaves per plant: the 200 Hildersham seedlings (frm parents f Lat. 52 N.) had 17.9^0.38 (cf. lcality 8, 20.0^1.87 in Table i). It must be remembered, f curse, as in all geneclgical experiments with grse, that distributins have ften been artificially effected. Sme ppulatins may nt have been established lng enugh t have reached a reasnable equilibrium with the lcal climate. Temperature and gentypic selectin Researches in New Zealand (e.g. Millener, 1947) have shwn hw impressively mean-minimum, and extreme-minimum, temperature figures can be made t accunt fr the facts f distributin f certain species there. While the seedling ectypes listed in Table i were grwing, therefre. New Zealand and Hildersham seedling ppulatins (amng thers) were raised in the glass-huse. They were then hardened ff utdrs fr nearly a further mnth. The New Zealand seedlings were then sme 12 cm high, and nearly all had sme aduh leaves. The mean height f the Hildersham plants was but 8 cm; and relatively few had left the juvenile stage. As has already been explained, Hildersham- seedlings typically prduce many tnre juvenile leaves than d New Zealand nes. The bxes f seedlings were then put in a chamber (Lw Temperature Statin, Cambridge) at 1 C fr 16 hurs, and then put ut in the pen again. There, they were

Day-length and Ulex eurpaeus 121 -H -H > q 00 00 ^ r^ ( 6 MM, -H -H -H t -H +1 N q d d -H M 9 1 CT\ r g^ r^-q r^v " 00 '^ " " H +1 -H H +1 -H +1 N c^ M -i" c> " i-i '> N N " q d -H CO M ri 6 13.5 1) ON M M N " 4-1) bj) m ^ > S.S.^ be C U 2 <d Wen «*^ a " 4-. O D Z :s. 1 V Z ^ S fa -Igathered seed c mm a ill Q c C 0 1 U Ihead X erse1 n, n rset B '^. u T.3 ffl 8 -B ill Wffl.s\ -d "a ^ ^ ^. -f 6 rt i~ + l

122 L. H. MiLLENER kept under bservatin fr 3 days. The abve treatments were tben repeated, but with successively lwer chamber temperatures: 1 C, 3 C, 7 ^ ^^^ ^^ Reactin at 7" C was quite nticeable, and at 10" C marked. Table 2 shws that the mre pen, mre rapidly-grwing N.Z. seedlings were less able t withstand the alternate freezing and thawing than the mre cmpact, slwer-grwing Hildersham nes. Mrever, all the N.Z. plants killed utright were f the tallest kind, with relatively few juvenile leaves. The mst resistant, n the ther hand, were f mre cmpact grwth, like the typical Hildersham seedling, and with relatively numerus juvenile leaves. Table 2. The end effects f sttccessively lwer temperatttreperids f lb hurs each {minimum 10 C) n tw ectypes f Ulex eurpaeus N.Z. ectype Hildersham ectype N. f plants initially 28 28 N. killed by treatment 7 3 N. cntinuing grwth frm riginal stem apex 5 N. cntinuing grwth frm lateral buds 21 25 The experiment can be regarded as f preliminary nature nly. T carry ut such a research satisfactrily many ppulatins f each type, staggered in develpment t varius degrees, wuld be needed; fr if the pst-juvenile phase f elngatin render the plant mst susceptible t frst damage, as may well be, the different types f seedling under experiment wuld have t be cmpared at similar levels f develpment, i.e. at similar physilgical, nt merely chrnlgical ages. Even s, the results make it pssible t suppse, tentatively, that the seedling type develping a relatively large number f juvenile leaves and passing thrugh the stemextensin phase relatively gradually, can the better withstand peridic lw temperatures such as are met with, in winter, in Eurpe. Ppulatins f grse seedlings are ntriusly hetergeneus (heterzygus) and, furthermre, the plants are sensitive, in fact diiterentially sensitive, t cld. Is it pssible that the mrphlgically varius, adult grse ppulatins in different parts f Britain have been, t sme degree, determined thrugh gentypic selectin, by diverse cld climates, frm hetergeneus seedling ppulatins? Such a prpsitin might explain beautifully the seriatin f the transitin numbers frm latitudinal grup t grup in Table i; fr just as the lcality is farther and farther nrth, s is there an increase in the mean number f juvenile leaves per plant. Hwever, the meterlgical data in Table i shws that the facts are nt in keeping with such a thery. There is n crrelatin between latitude and temperature. The chief mderating influence is the sea, particularly the warm Atlantic waters. Cambridgeshire, nt nrthern Sctland, has the mst severe temperature climate as far as minima (bth mean and extreme) are cncerned. Salisbury (1932) has emphasized the cntinental qualities f this climate. Day-length and gentypic selectin Temperature gradatin having been dismissed as the factr gverning the psitive crrelatin between the extent f juvenile develpment in grse and increase in latitude, it wuld seem that day-length is the nly envirnmental factr changing regularly with latitude that might explain the bserved seedling variatins.

Day-length and Ulex eurpaeus 123 Several demnstratins have nw been given f ectypic differentiatin in ther wdy species (e.g. Ppidus Pauley and Perry, 1954) that can be crrelated with daylength variatins ccurring ver the natural gegraphic range f the species. In grse itself seedling develpment has been shwn t be greatly affected by artificially cntrlled variatins in day-length (Millener, in press). Day-length cntrl is als exercised, f curse, by the cycle f the seasns. Nwhere in the extensive literature n Ulex des there appear t be reference t changes in seedling mrphlgy with the seasns. There is little dubt, hwever, that Bdle (1914) wuld have recrded the phenmenn had he persisted a little lnger with his experiments fr seasnal cntrl ver the mrphlgy f the grse seedling is a reality. JAN. FEB. MAR. APRIL MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. 5 = 16-7 t O-94 5-12-4 ± O-I7 Seedlings grwn in J.I.C.N.l 5 = 14-6 ± O-57 ; = 3-8 ± O-64 t X = 19-7 + -23 6-4±O-3- = 14-3+0-=9 -IO-66± 0-71 Seedlings grwn in Hildersham sil X = 13-7 ± 0-80 1 5E=l8-2±O-46 x=26-l ± 1-22 Fig. I. The march f mean transitin numbers (.v, r the mean numbers f juvenile leaves prduced per seedling ppulatin) in grse grwn thrugh the seasns in tw different media. Fig. I shws the march f the mean transitin numbers f seedlings f the N.Z. ectype, swn in tw different media, thrugh the seasns. 'Jhn Innes Cmpst', N. i, is the mdified Jhn Innes mixture already described, save that the peat is Irish Sphagnum peat nly. 'Hildersham sil' is a rather pr, acid (phs), sandy sil frm Hildersham, near Cambridge, in which grse grws naturally and well. There were 100 seedlings in each ppulatin save fr thse swn in Hildersham sil in February and August, in each f which there were 200, and that swn in cmpst in May, previusly mentined in this paper, in which there were 400 a ttal f 1600 plants altgether. A great advantage f using the transitin number as the key attribute t measure prgress in the develpment f seedlings, and perhaps f their physilgical age, is that the analysis date is immaterial, prvided the plants have all becme 'adult'. This state was reached within 3 mnths by all the ppulatins represented in Fig. i, save fr the

124 L. H. MiLLENER September-swn ne which had t be ver-wintered and analysed the fllwing spring after it had cmpleted juvenile develpment. All the ppulatins, except this ne, are represented arbitrarily as having had grwing perids f 3 mnths' duratin. Essentially the same seasnal pattern f drift in the transitin number is shwn by the cmpst-grwn and sand-grwn ppulatins: but sand-grwn plants develp regularly fewer juvenile leaves prvided the swing dates are cmparable a cnfirmatin f Bdle's (1914) tentative cnclusins frm grse ppulatins grwn n sand and sil. Fig. 2. Day-length curves thrugh the year fr three statins, shwing als the general relatinship f.-v values t changing day-length thrugh the year. The ver-wintered ppulatin has been excluded. If the ppulatins be cnsidered in turn, it can be prven that the transitin numbers f any tw successive ppulatins are significantly different frm ne anther, excepting nly the tw cmpst-grwn sets swn in June which can be cnsidered as replicates. The curve f transitin numbers thrugh the year is hllw (Fig. 2). The values fall thrugh spring t a minimum in summer, and then rise nce mre t an autumn value higher than that f spring. This is exactly the behaviur that might have been predicted frm experiments n grse seedhng ppulatins grwn in a cntrlled-cnditins cabinet under different day-lengths (Millener, in press). Cntrlled-cnditins experiments (Millener, 1952) have indicated, t, that the very different hght-level climates under which the ppulatins were grwn thrugh the year are prbably nt f cardinal imprtance in cntrllmg the extent f the juvenile phase f leaf frm - althugh spme, leaf size, mternde length, and sme ther characters, are very dependent n light. Similar experiments (Millener, 1952) have made it dubtful whether temperature changes, per se, play a dminating rle. If, in the seasnal experiments under discussin,

Day-length and Ulex eurpaeus 125 relatively lwer teniperatures had brught abut the higher transitin numbers f the late-seasn seedlings as cmpared with the summer-grwn nes, shuld nt the Februaryswn ppulatins have had still mre, nt fewer, juvenile leaves? Minimum, mean and maximum temperature figures fr February-April are all significantly lwer than fr August-Octber. Day-length itself then wuld seem t be the key factr in the seasnal cntrl f leaf frm in grse. The picture, hwever, can hardly be as simple as the abve arguments may seem t have implied. Beynd dubt, many factrs varying with the seasns must perate tgether t prduce the kind f changes that have been measured. There is evidence, fr example (Millener, 1952), that high humidities (which can infiuence significantly the juvenile-adult balance in grse) may affect leaf frmatin mre prfundly under shrt-day than under lng-day cnditins. Whatever the reasns, the late autumn effect is very striking. The even mre remarkable influence f the ver-wintering f partly develped seedlings n leaf develpment (.v = 26.1 fr the Hildersham sil plants swn in September) culd cnceivably be due mainly t a day-length effect: but it is very likely that an abnrmally high rt/sht rati, is cncerned as well. Other experiments have shwn that ver-wintered grse seedlings and plants struck late in the seasn frm cuttings, develp a remarkable mass f fibrus rts by the time spring grwth f the fllwing seasn is well under way. The bservatins made by Schuepp and Chang (1938) n Sinapis alba and by Ashby and Wangermann n Ipmea piirpurea (1950) have led t the suggestin that day-length at the time f germinatin is the perative factr in setting the pace fr leaf prductin. Fr each week later that seed f Ipmea is swn (between 5 April and 17 May), the mean time interval between successive leaves 1-9 is reduced by abut half a day, and the mean maximum gr\\1:h rate is increased abut 40%. These are remarkable respnses since day-length nly increased frm abut 13I hurs at the first swing t 15 hurs at the last, i.e. abut 15 minutes frm swing date t swing date. In the grse experiments the day-length difference between ne swing and the next varied frm sme ii hurs t ver 5 hurs. The irregular intervals between swings are due t the fact that the experiments were nt designed, in the first instance, fr daylength researches. On the basis f the Ipmea experiments, then, day-length differences between swings might well be expected t exercise a decisive infiuence ver leaf develpment in grse the actual leaf sequence in this case being affected, as well as rate f leaf prductin, etc. The data are nt cnsistent, hwever, with such an extended hypthesis (Table 3). Cntrlled day-length experiments n grse seedlings (Millener, in press) have invariably shwn a negative crrelatin between day-length and transitin number. The swings in the shrt days f February, therefre, shuld have prduced seedlings with a far higher transitin number than the swings in the lng days f early August. But this is nt s. Other discrepancies may be seen amng the values in the table. If, hwever, day-length acts mainly when the seedlings are at the height f juvenile leaf prductin abut the middle f the 3-mnth grwth perids shwn in Fig. i then negative crrelatins betw^een day-lengths and transitin numbers, thse f the August-swn ppulatins alne excepted, are very satisfactrily shwn. The juvenile leaf numbers f the August-swn plants are a little high because f the apparent 'autumn effect' already mentined. The values in Table 3 cannt, unfrtunately, be analysed statistically. It shuld be pssible t test even mre decisively the tw pssibilities. Seed wuld be

126 L. H. MILLENER swn 2-3 mnths befre the lngest day s that the just-'adult' ppulatin culd be analysed then. On that day a secnd swing wuld be made, the plants being analysed after the same grwth perid. The ppulatins wuld, f curse, experience the same day-lengths, nly 'in reverse': day-lengths at the mid-perids f grwth wuld be essentially the same. Table 3. Mean transitin numbers (x) in relatin t apprximately day-lengths {sunrise t sunset) at different stages f develpment J.l.C, N. I X Day-length at germinatin Day-length at mid-perid f grwth Feb. swing 16.7 9 hr 56 n-iin 13 hr I min May swing 12.4 15 hr 4 min 16 hr 33 min June swings 14.2 (pled mean) 16 hr 37 min 14 hr 53 min Aug. swing 19.7 15 hr 9 min 13 hr 4 min Hildersham sil X Feb. swing Mar. swing May swing July swing Aug. swing 16.4 14.3 10.6 13.7 18.2 Day-length at germinatin 9 hr 6 min 10 hr 48 mm 15 hr 5 min 16 hr 37 min 14 hr 45 mm Day-length at mid-perid f grwth 11 hr 54 min 14 hr 16 hr 33 min 14 hr 28 min 12 hr 51 mm CONCLUSION Changing day-length thrugh the seasns is clearly imprtant fr the vegetative develpment f the grse seedling; and it is equally certain that n envirnmental factr besides the length f day can be invked t explain {a) the mre r less predictable variatin prduced in the leaf sequence f a given seedling ppulatin by grwing it in ne latitude instead f anther (Millener, 1952) and {b) certain 'fixed', r genetic, variatins between ppulatins naturally grwing in different latitudes. Table 4. Apprximate day-lengths {sunrise t sunset) in hurs and minutes thrugh the year at three statins Jan. I Feb. i Mar. i Apr. 1 May i June i July i Aug. i Sept. i Oct. i Nv. i Dec. i Exeter 8.01 9.13 10.50 13.01 14-50 16.14 16.28 15.25 13.30 11.36 9.44 8.17 (50" 44' N) Cambridge 7.52 9.06 10.48 13.01 14.57 16.24 16.39 15-33 i3-34 ii-34 9-39 8.08 (52" 10' N) Inverness 6.53 8.26 10.32 13.13 15.33 17.26 17.47 16.20 13.48 11.30 9.07 7.14 (57 33'N) It wuld perhaps seem natural t seek an explanatin f the ectypic variatins in leaf sequence that are bund up with latitude by studying the behaviur f ppulatins grwn in ne place but at different times (i.e. under different day-lengths) thrughut the seasns. A critical examinatin f the facts available, hwever, makes it certain that daylength diflferences frm latitude t latitude are nt cnsistent with the kinds f leaf sec[uences that might be thught, frm a study f the ectypes grwn tgether under unifrm cnditins, as typical f thse latitudes. Table 4 and Fig. 2 shw that at Inverness, the mst nrtherly statin, where seedlings are inherently mre juvenile (i.e. are mre f the shrt-day type) than farther suth, the

Day-length and Ulex eurpaeus 127 days are actually lnger frm the latter half f March until twards the clse f September than at the mst sutherly statin (Exeter). These 6 mnths cver mst f the grwing seasn fr grse. They cver, even, mst f the perid ver which seed germinatin is pssible, s that even if day-length at the time f germinatin did set the pattern fr leaf develpment in grse (and it apparently des nt) shrt days wuld nt usually accmpany seed germinatin in the nrth, and lng days in the suth. The seeds f grse, in fact, althugh shed at the height f summer, d nt germinate habitually at ne particular seasn. If they did, day-length differences in the fllwing mnths between ne latitude and anther might be felt by the seedlings. Because f their 'impermeable' testas, hwever, seeds tend t germinate randmly right thrugh the grwing seasn. Finally, it is difficult t see hw day-length variatins, even if they were significantlytied t seedling behaviurs at the different statins, culd directly 'select' seedlings as culd, say, severe temperature variatins. Pauley and Perry (1954), fr example, were able t see day-length functining as the timing device which enables height-grwth cessatin in Ppidus t be crrelated with the length f the frst-free seasn in varius native habitats. In grse such hereditary mdificatins in the leaf sequences f seedlings as are due t the influence f day-length must have their genesis in the flwer-buds f the parent plants. These buds are typically prduced in Britain in autumn and early winter and meisis ccurs when the buds are less than 3 mm in length, i.e. lng befre the flwers pen in spring (r even, smetimes, in winter). This means, f curse, that shrt-day cnditins almst invariably attend thse cell and nuclear prcesses where hereditary changes f sme kind mst usually ccur. The influence f day-length upn grse is s prfund that ne may suspect this factr t perate thrugh an integrated cell chemistry mechanism that prbably affects all grwth prcesses. It is nt t difficult t imagine that prcesses in the hereditary material itself might be mdified t different degrees by the dissimilar shrt-day climates f different latitudes. The changes that have been induced in grse, presumably ver a lng time, thrugh these quite unknwn mdificatins t unknwn prcesses, have led t the prductin f seedlings that have essentially the qualities that tend t be impressed by the particular latitude upn any seedlings grwn there. Relatively lw latitudes are related t seedlings with few juvenile leaves: relatively high latitudes engender the develpment f many juvenile leaves. REFERENCES ASHBY E & WANGERMANN, E. (1950). Studies in the mrphgenesis f leaves. IV. Further bservatins n area, cell size and cell number f leaves f Ipmea in relatin t their psitin n the sht. BODLErL. A.''{i9tl)^On the trifliate and ther leaves f the grse (Ulex eurpaeus L.). Ann. BL, Land., LAWRENCEJW. C. J. & NEWELL, J. (1942). Seed and Ptting Cmpsts 2nd edn. Lndn. MILLENER, L. H. (1947). A study ientelea arbrescens R.Br. ('Whau'). Part i - Eclgy. Trans. Ry. Sc. Mii.^ifzR.'^t. il.\igs2). Experimental studies n the grwth frms f the British species f Ulex L. Ph.D.' thesis, Cambridge University Library...,. rr; ^ T -ru MILLENER, L. H! (1961). Day-length as related t vegetative develpment in Ulex eurpaeus L. i. Ihe experimental apprach. ive^ P/zyiZ., 60, 339. r> 1 -^ PAULEY, S. S. & PERHY, T. O. (1954). Ectypic variatin f the phtperidic respnse in Ppulus. J. S.LISIURY'^E'j (i9"'5. The E^ast Anglian flra. Trans. Nrf. and Nrw.Nat. Sc, 13, 191. SCHOEPP, b & CHANG, C. Y. (1938). Einfiuss der Lichtintensitat des Sprsses vn Sinapis alba. Verh. WAGER!'??.- SWf tsei^'at'ins n the mrphlgy f speces f the genus Ulex. Int. J. Mrc. Nat. Sc, 3rd Ser., 7, 36. B N.P.