Germination behaviour of the seeds of seven New Zealand vine species

Transcription

1 New Zealand Jurnal f Btany ISSN: (Print) (Online) Jurnal hmepage: Germinatin behaviur f the seeds f seven New Zealand vine species C. J. Burrws T cite this article: C. J. Burrws (16) Germinatin behaviur f the seeds f seven New Zealand vine species, New Zealand Jurnal f Btany, 34:1, 3-102, DOI: / T link t this article: Published nline: 07 Dec Submit yur article t this jurnal Article views: 165 View related articles Citing articles: 14 View citing articles Full Terms & Cnditins f access and use can be fund at Dwnlad by: [ ] Date: 01 December 2017, At: 11:41

2 New Zealand Jurnal f Btany, 16, Vl. 34: /6/3401)3 $2.50/0 The Ryal Sciety f New Zealand 16 3 Germinatin behaviur f the seeds f seven New Zealand vine species Dwnladed by [ ] at 11:41 01 December 2017 C. J. BURROWS Department f Plant and Micrbial Sciences University f Canterbury Private Bag 4800 Christchurch, New Zealand Abstract Germinatin rates, percentage germinatin success, and phenmena related t germinatin delay were determined fr seeds frm freshly cllected fruit (summer-autumn) f Calystegia tugurirum, Clematis fetida, Muehlenbeckia australis, Parsnsia heterphylla, Ripgnum scandens, Rubus cissides, and Tetrapathaea tetrandra. The experimental treatments simulated natural cnditins that the seeds might experience after dispersal. In the standard treatment (seeds cleaned, mist, well-lit) seeds f each species behaved differently. Ripgnum seeds germinated thrugh autumn and winter; in late-cllected fruit sme had germinated viviparusly. The large seeds have a very thin testa and are killed by drying. Sme Muehlenbeckia seeds germinated in autumn, the rest did nt d s until early spring. Clematis and Tetrapathaea seeds did nt start germinating until late winter (when temperatures were rising) and all viable seeds had germinated within tw mnths. A few Rubus and Parsnsia seeds germinated, respectively, in autumn and spring. Germinatin f the remaining seeds ccurred at a slw, steady rate fr Rubus, fr tw years, and during warm perids fr Parsnsia, fr three years. Spradic germinatin f hard-cated Calystegia seeds ccurred ver five years. The amunt f germinatin success was high (8-8%) fr Clematis, Ripgnum, Muehlenbeckia, and Tetrapathaea, but nly 18, 73, and 82%, respectively, fr Calystegia, Parsnsia, and Rubus. B5010 Received 28 February 15; accepted 13 August 15 In the dark treatment mst Tetrapathaea and Parsnsia seeds germinated, but nly mderate numbers f Ripgnum, Muehlenbeckia, and Rubus seeds and nne f Clematis. In the sil treatment germinatin success was lwer than in the standard treatment and germinatin rates were slwer. The same applied t seeds in the in-fruit treatment fr Muehlenbeckia, Parsnsia, and Rubus. N radicles emerged in the in-fruit treatment fr Ripgnum, but the embrys f many develped, viviparusly, t the immediate pre-emergence stage. Ripgnum, Calystegia, Clematis, Muehlenbeckia, and Tetrapathaea can all maintain shrt-term seed banks, in fruit, n the parents (fr tw mnths r mre). Rubus, Parsnsia, and Calystegia seeds can persist in sil seed banks fr mre than a year. Ripgnum des nt maintain sil seed banks. It is nt knwn whether seeds f Clematis and Tetrapathaea can d s. Generally the seed germinatin phenmena accrd well with the habitat preferences f the respective species. Keywrds seeds; dry fruit; fleshy fruit; germinatin delays; different patterns; vivipary; seed banks; envirnmental crrelatins INTRODUCTION The wrk described here cntinues a series arising frm research n the germinatin f seeds in New Zealand lwland frest wdy plant species, in cnditins similar t thse that the seeds wuld experience in nature (Burrws 15a, 15b, 15c). Seven vine species were studied. Very little appears t have been published n the germinatin f seeds f New Zealand vines hithert (cf. Macmillan (172) (Ripgnum); Cartman (183) (Clematis)). Cnsidering the imprtance f vines in the flra and vegetatin (Dawsn 188; Burrws 14), this is surprising. The aim f this accunt is t prvide basic infrmatin n germinatin rates, numbers f seeds which finally germinate, and features f the germinatin

3 4 New Zealand Jurnal f Btany, 16, Vl. 34 Dwnladed by [ ] at 11:41 01 December 2017 delay systems fr seeds cllected fresh frm the parent plants. The cnceptual framewrk fr the study is that utlined in Burrws (18, 14). Plant nmenclature fllws Allan (161) and Mre & Edgar (170). The individual disseminules are all referred t as seeds, althugh three include vary tissue, in the frm f the entire vary wall (Clematis, Muehlenbeckia), r as endcarp, the tugh inner part f the vary wall (Rubus). The species tested Calystegia tugurirum (Frst.f.) R.Br. ex Hk.f. (Cnvlvulaceae) A twining climber, evergreen in mild climates and summer-green in clder climates where the species may be suffrutescent, grwing stems anew each year frm a wdy base. The thin stems may be at least 5 m lng. The bisexual flwers are brne singly in the axils f leaves. The fruit are indehiscent, singlelculed capsules, each cntaining fur range seeds, with basal placentatin. Ripe seeds have very thick, hard cats, which are water-impermeable. The species is widespread frm the cast t upland lcatins in scrub, at frest margins, r in lw frest, where it reaches the canpy. Clematisfetida Raul (Ranunculaceae) An evergreen, diecius vine which climbs with the aid f twining petiles. The ften tangled stems may be at least 8 m lng. The flwers are brne in small panicles in the axils f leaves. Each female flwer has a large number f achenes, each with a lng plumed tail, the style. The tails usually curl ver t frm a glbse seed head. The species is cmmn at lwland frest margins, r in lw frest, where it reaches the canpy. Muehlenbeckia australis (Frst.f.) Meissner (Plygnaceae) An evergreen, diecius, twining climber ften with multiple, tangled stems up t abut 10 m lng. The female flwers are brne in large axillary panicles subtended by sheathing stipules. The fruit, a nutlike achene, is surrunded at its base either by dry green, r fleshy white perianth lbes. The species is widespread frm the cast t inland and upland lcatins, in scrub, at frest margins, and in frest, where it reaches the canpy. Parsnsia heterphylla A.Cunn. (Apcynaceae) An evergreen, twining climber with stems at least 10 m lng. The panicle-like inflrescences f bisexual flwers are terminal n branches, r axillary. Each fruit is a lng, dehiscent pd, with tw lcules. The numerus seeds each have a tuft f hair at the distal end. The species is abundant at lwland frest margins, r in frest, where it reaches the canpy. Ripgnum scandens J.R.Frst. & G.Frst. (Smilacaceae) An evergreen, twining climber. The multiple, tangled stems may be at least 10 m lng. The bisexual flwers ccur in small t large terminal r axillary racemes. The fruit are bright red berries with juicy pericarps. Each cntains ne, tw, r smetimes three large, thin-walled seeds. The species lives in dense lwland frest, where it reaches the canpy. Smetimes it ccurs at frest margins. Rubus cissides A.Cunn. (Rsaceae) An evergreen, diecius climber which ascends with the aid f its recurved thrns and by twining. Its ften multiple, tangled stems may be at least 10 m lng. They scramble ver the grund as well as n shrubs and trees. The female inflrescences are large panicles, brne in the axils f leaves. The aggregate fruit are fused clusters f drupelets with fleshy, reddish pericarp tissue. Each has ne seed surrunded by endcarp. Drupelets are clustered tightly arund a receptacle. The species is widespread frm the cast t inland and upland lcalities at frest margins. It may be abundant in scrub, at frest margins, r in the frest subcanpy, but it als cmmnly reaches the canpy. Tetrapathaea tetrandra (Sl.) Cheeseman (Passiflraceae) An evergreen, diecius vine which climbs with the aid f tendrils. The ften multiple, tangled stems may be up t 10 m lng. The flwers are brne singly, r in few-flwered cymes, in the axils f leaves. The fruit are large, bright range, dry-walled, threelculed berries with parietal placentatin. The seeds are each surrunded by a sac f succulent red flesh. The species ccurs at lwland frest margins, r in frest, where it reaches the canpy. Seed dispersal The seeds f Clematis and Parsnsia are dispersed by wind, thse f Calystegia by gravity r pssibly in streams. Seeds f the ther fur species, invested by fleshy pericarp tissue (Ripgnum, Rubus), mescarp (Tetrapathaea), r perianth (Muehlenbeckia), are dispersed by birds, except that sme Muehlenbeckia seeds, surrunded by dry perianths, prbably disperse by simply falling.

4 Burrwserminatin behaviur f seven NZ vines 5 A 25- u' 20- EE w E : IO- 5 0 O O.. F. M. A. M. J.. JAS. N J. F. M. A. M. J... JA ON ' 6 ' O F A.. MJ.. J. A. S. O. N J F ' I 18 I 10 I 11 I Fig. 1 Mean mnthly maximum and minimum temperatures 18-2 fr the Canterbury Btanical Gardens site (data frm the New Zealand Meterlgical Service). Dwnladed by [ ] at 11:41 01 December 2017 Kereru (Hemiphaga nvae-seelandiae) were seen eating Ripgnum fruit; bellbirds (Anthrnis melanura) ate Rubus fruit; blackbirds (Turdus merula) ate Ripgnum and Tetrapathaea fruit; and silvereyes (Zsterps lateralis) ate Muehlenbeckia, Rubus, and Tetrapathaea fruit. The Tetrapathaea fruit walls are trn pen by the birds, which swallw the internal flesh cmplete with seeds. METHODS The general methdlgy was the same as that utlined fr experiments described by Burrws (15a). Ripe fruit were gathered, mainly frm lcatins in Canterbury (Table 1), and the seeds remved fr all but ne treatment, by maceratin and decantatin. Mdular replicates f 25 seeds were used fr each test. The treatments were: standard (pericarps remved, seeds washed and maintained in petri dishes n filter paper, in wet, well lit cnditins, 4 replicates); dark (cnditins the same, except that petri dishes were wrapped in aluminium fil and examined peridically in a darkrm, under a "Kdak" phtgraphic safelight, 2 replicates); sil (cnditins the same as fr the standard except that seeds were placed n mist sil in small plastic trays, 2 replicates); in-fruit (cnditins the same as fr the standard except that pericarps were left arund the seeds, 2 replicates). The in-fruit treatment was nt dne fr Tetrapathaea. Clematis and Muehlenbeckia seeds are tightly invested by varian tissue which was nt remved. Muehlenbeckia seeds were left in perianth tissue fr this test. At the beginning f the experiments tetrazlium tests fr ptential germinability were dne n 25 fresh seeds f each species. A small supplementary experiment with 10 Calystegia seeds frm the same lcatin (Kaituna Valley), carried ut in March 11, cnsisted f nicking the seed cat with a nail file t make it permeable t water. The standard, in-fruit, and sil treatments were checked at abut weekly intervals and the dark treatment abut nce mnthly. The first signs f germinatin in the seeds were: Calystegia, swelling f the seed t abut twice its dry vlume as imbibitin takes place, then splitting f the seed cat and radicle emergence; Clematis, Muehlenbeckia, Parsnsia, Rubus, and Tetrapathaea, splitting f the seed cat and radicle emergence; Ripgnum, appearance f the green embry n the uter wall f the seed. The experiments were maintained in an unheated, partially-shaded glasshuse with autmatic watering, at the University f Canterbury Department f Plant and Micrbial Sciences btanical garden area (cf. Burrws 15a). The temperature mnitring was dne with maximum and minimum thermmeters and a Casella mercury-in-steel thermgraph. Hwever, the mre cmplete Christchurch Btanical Gardens climatic recrd is used (Fig. 1) as a prxy fr the University glasshuse recrds (which it resembles clsely) (Burrws 15a). As the within-treatment variability f the results (final germinatin success) is lw fr each species, n detailed statistical treatment was dne. Graphical representatin f the results is cnsidered adequate. RESULTS The results are presented in Table 2 and Fig Calystegia is a special case which is cnsidered separately. In the standard treatment individuals in the

5 6 New Zealand Jurnal f Btany, 16, Vl. 34 \0 0\ (1) Z :::: N (1) ::I Q.. E: c 3 E:..-..., t " 0 0 ::I \0 \0,0\ 2- < w.". Dwnladed by [ ] at 11:41 01 December 2017 Table 1 Fruit and seed prperties. Cunts and measurements were dne n randm samples f freshly cllected fruit and thrughly air-dried seeds frm the cllectin lcalities; unless therwise indicated ten individuals were cunted and measured. The values are means and standard deviatins. N errr values are prvided fr the mean seed weights which were btained by weighing a given number f seeds. The seeds r fruit f Clematis, Parsnsia, and Tetrapathaea are flattened; the narrwest dimensin was nt measured. Seed cunts were dne n Clematis heads and Rubus aggregate fruit. The value in parentheses fr Clematis is numbers f heads per panicle. Cmparative measurements f wet and dry weights were dne n a set f 10 Ripgnum seeds. Z =,.. _.= (D e",,.- ", "6N e.= N an Cllectin date; N. f Ripe fruit Grid reference fruit per length N. f Taxn; (NZMS 260); infruct- f main Ripe fruit full seeds Seed length Seed width Mean seed Lcatin Altitude (m) escence axis (mm) diameter (mm) per fruit (mm) (mm) weight (g) 11 7c.r r D OQ 3 g Calystegia tugurirum 10 Mar ± ± ± ± ± t'q 4-1, c r O D I " -I cq c l-i,,d r eq Kaituna Valley M36/84181 western Banks Peninsula 1 Clematis jetida 15 Feb 0 31.± ± ± 0.2 see fruit dimensins e,i Ahuriri, Prt Hills M (13.7 ± 4.6) western Banks Peninsula 420 Muehlenbeckia australis 15 Feb ± ± ± 0.2 see fruit dimensins Crater Rim, Prt Hills M western Banks Peninsula 318 Parsnsia heterphylla 15 Apr ± ±.2 4. ± ± ± ± Ahuriri, Prt Hills M western Banks Peninsula 420 Ripgnum scandens 8 Mar ± ± ± 1. IA± ± ± 0.6 weto Peraki Saddle Reserve N37/8704 dryo.131 eastern Banks Peninsula Rubus cissides 4 Jan ± ± ± ± ± ± b 4-1 +I 4-1 -t-1, " c5 e t c'q -t-i H - i'-,1 t " r'q - c-q r t'- m.,4 " cq Ahuriri, Prt Hills M western Banks Peninsula 420 Tetrapathaea tetrandra 1 Apr ± ± 0.8 loa ± ± ± Riccartn Bush M Christchurch.5 et t " < -. <

6 Burrwserminatin Burrws--Germinatin behaviur f seven NZ vines 7 Dwnladed by [ ] at 11:41 01 December 2017 C.S c ' Oll (/] M e,i aj :c E- C ' Oll en en 4- C/....n E ;::l Z m '0 (/].;:; <I::.5 -'"... CI... C czl ;::l - ;.=.n E OJ I. 0 i;jf-> ) b- f- "::? '::? "::? '::? r "::? '::? en x en I I I, t; OllC 0 ;; v >'"0., c E'- -CiE... u -a.!.c: E ;;.g {/) '. CI ";:: S a Oll C. czl C f- OJ u C " ON\OO"<t O c 0, '<S 00 S \ONO ("'1\0 01 )00 t-'1"01 t-"<tn t-'1" t-"<t 0'-:"<1'00'-: 0'-:'<1'00'-: [ )0001 )0 0r--:0i"<1'0 0r--:00f'i'<1'0 --N-N \ t- «") "E -.::t: t-:;.{g- C C: V:.NO\ ti N - 0\ C NONON t-"<t00"<to\ "<ti )N"<t '1"1 )01'1" O-:M'-; +q ) «") ",,; --N-N 00 O\t-«")N 00 0\ t-o\00o\ Ot.r)O\\OV)V) Ot.r)O\\Ot.r)V) 6""";6N q q b If) I ) tr)t.r)if)t.r)v)'f) NNN... '<1' "<1' '"'!'"'!'"'!'<1'0'<1'... "<t O"<t 01'1" N"<tNNN NN- "0«") O\ - u 00\ u _ -1 )-\0 6'0 :;; 8 >. g.n 00 O\1 )"<t00t 0\ I ) ;::: t: :;:.S!.0 '" c3 '::j"r')-v)m 000\\- 0\ 0\ 00 0\ 00 0\.n.n... c>. 3-.;g "'"''""'"''''''''..., ;?,NI )_,,- «") ;:: ;::::::::: =:: =:: ;:J,{l ;:: 2 =:: '" '" '".t1;;;:sl%:l -<:: ;::: 0....dil:! w O();:J ;::: '" l::....-::! \..)...c::.::c::.::e-:; sets f seeds fr Ripgnum and Muehlenbeckia, cllected, respectively, in early autumn (March) and late summer (February), germinated at a slw, steady rate in autumn (Fig. 2, 3). Mre Ripgnum seeds germinated thrugh winter and the last few in August. N Muehlenbeckia seeds germinated in win- winter and the final flush ccurred in early spring. Rip- Ripgnum seeds in the dark and n sil were slw t germinate cmpared with thse in standard cndi- cnditins. In the sense that there was n radicle emergence, nne f the Ripgnum in-fruit treatment seeds ger- germinated. Hwever, 5% f a set f seeds cllected frm parent plants in the same Banks Peninsula l- lcality in late autumn were fund t have germinated in the fruit (i.e., were viviparus) (Burrws 13). A check f the in-fruit treatment seeds shwed that a similar prprtin had als germinated vivipar- viviparusly. In these the green and undevelped seedling was evident as a "windw" n the seed wall. Amng the late-cllected seeds frm the field, hwever, a few f the germinated seeds had the radicle emerg- emerging abut 0.3 mm. Vivipary has subsequently been bserved in late-cllected Ripgnum seeds frm ther lcatins (c. (C. 1. J. Burrws unpubl. unpubl, data). Seeds frm Ahuriri Valley Bush, western Banks Peninsula (June 14) and frm the Wainihinihi River, Taramakau Valley, Westland (December 14) had all germinated. Mst were at the "green windw" stage but abut 5% in each case had prtruding radicles. Grwth f emerged Ripgnum seedlings is slw; they remain attached t the large seed fr several mnths. A test f viability after 10 seeds had been air-dried fr 3 weeks shwed that all had died. Macmillan {I (172) als recrded the death f dried seeds. Seeds f Muehlenbeckia in the dark and sil treat treat- ments germinated mre slwly than in the standard, and the final amunt f success was markedly lwer. Hwever, althugh seeds in the in-fruit test were slw t begin germinatin the final success was nly slightly lwer than in the standard. Clematis and Tetrapathaea seeds, cllected in late summer and autumn, respectively, did nt underg a substantial flush f germinatin in the standard treatment until late winter (when temperatures were rising, Fig. I) 1) (Fig. 4, 5). In particular there was a rapid burst f germinatin in Tetrapathaea. Seeds f this species in the dark treatment respnded very similarly, but the germinatin rate fr seeds n sil was slw and final success was nly 80%. N Clematis seeds germinated in the dark and there was pr germinatin n sil fr this species. Tara-

7 8 New Zealand Jurnal f Btany, 16, Vl O) J' r S w Treatment standard dark xxx in fruit & sil l?/pgnum scandens Peraki Saddle, Banks Peninsula MAR I APR I MAY I dun I JUL I AUG I SEPT Fig Germinatin curves fr Ripgnum scandens seeds. Dwnladed by [ ] at 11:41 01 December z 85 Muehlenbeclda australis Prt Hills, Banks Peninsula 0 0 x Z x 0 ll ' I 1 I I I I I I I 1 I I I I I I IO I IO IO O IO IO FEB MAR APR MAY JUN AUG SEPT 10 Clematis fet/da Prt Hills, Banks Peninsula Fig. 3 Germinatin curves fr Muehlenbeckia australis seeds. (In-fruit treatment with achenes surrunded by perianth.) Symbls as in Fig. 2. Tetrapathaea te trandra Riccartn Bush, Christchurch 25- " 20-, t-,,- 88, 5,, E 0.,,./.,e A, A P 1 z z 30, zb 3', z 3b FEB JUL l AUG I SEPT 10 Fig. 4 Germinatin curves fr Clematisfetida seeds. (As the "seeds" are, strictly, achenes, the standard and in-fruit treatment are, in effect, the same.) Symbls as in Fig ", gg 0 82 t e ' 3', 2b 3',; 2' APR JUNI JUL I AUG I SEPT 11 Fig. 5 Germinatin curves fr Tetrapathaea tetrandra seeds. Symbls as in Fig. 2.

8 Burrwserminatin behaviur f seven NZ vines Prsnsi heferphyll Prt Hills, Banks Peninsul w Fig. 6 IO APR 0 A, 2'0 3, 2' SEPT OCT A 0,, -,,,....../,, 30 I lo NOV DEC JAN FEB l Germinatin curves fr Parsnsia heterphylla seeds. Symbls as in Fig. 2. Q I NOV 30' 3'' DEC JAN I i i 30 FEB 11 i 3O Dwnladed by [ ] at 11:41 01 December " 20- m -,w E O O Fig. 7 JAN Rubus ciss/des Prt Hills, Banks Peninsula 0,./%,., Q 0.'" I r i,, ; i,; ; ', 2? ',,, 18 Germinatin curves fr Rubus cissides seeds. Symbls as in Fig. 2. Fr all f the species cnsidered s far germinatin success in the standard treatment was high (8-8%). This cntrasts with the lwer degree f success fr the remaining three species (18-82%). In the standard treatment the first Parsnsia seeds, cllected in April, germinated in spring (September) (Fig. 6). Frm then n the rate was very slw until February f the fllwing year. A few mre seeds germinated in February the next year (final success 73%). Seeds in the dark, n sil, and in-fruit germinated at abut the same rate as fr the standard treatment seeds. Nne f these latter were maintained beynd the early part f the secnd year (10). The first f the standard treatment seeds f Rubus, cllected in January, germinated in March. Frm then n the germinatin pattern was similar t that fr Parsnsia but the last seeds did nt germinate (82% final success) until Octber 10, the secnd year after the start f the experiment (Fig. 7). Seeds in the dark and in-fruit germinated mre slwly, with lwer success, while seeds n sil were intermediate in germinatin rate and degree f success between the standard and the ther tw treatments. 20 SEPT DEC FEB APR AUG OCT ] 10 The mst spradic germinatin in the standard treatment and lwest final germinatin success was fr Calystegia seeds. In the first year f the experiment (18) nly 3 seeds germinated and the numbers fr successive years were: 10 (1); 11 (2); 12 (3); 13 (). This makes a ttal f 18 seeds (ut f 100) in 5 years. In the first tw years 6 seeds (ut f 50) germinated in the dark treatment, 2 n sil, and 1 in-fruit. The hard seed cats f this species are impervius t water and the barrier is nly slwly brken dwn, presumably by micrbial decay f the uter testa. The dark treatment success was cnsiderably higher than that fr the standard. Further experimentatin n this phenmenn is needed. In the supplementary experiment with Calystegia seeds with nicked cats the seeds were all swllen, t at least twice the riginal vlume, the day after being placed in experimental cnditins and the radicles (and in sme cases the whle seedlings) f all had emerged within 13 days. N seeds in an untreated cntrl set had germinated after 5 mnths. Swelling f the seeds ccurs thrugh absrptin f water by an inner seed cat.

9 100 New Zealand Jurnal f Btany, 16, Vl. 34 Dwnladed by [ ] at 11:41 01 December 2017 Tetrazlium tests Results frm the tests fr seed viability (Table 2) are generally clse t the actual germinatin success except fr Clematis, where a lwer scre was btained with tetrazlium, and Calystegia and Parsnsia where the tetrazlium result was higher. Fr Calystegia the discrepancy is explicable in terms f the lng delay in germinatin f mst seeds impsed by the hard seed cats. In Parsnsia there is als drawn-ut germinatin but sme seeds in the treatments die thrugh infectin by fungi r algae. DISCUSSION Germinatin patterns and germinatin delay Althugh each f the vines studied here may be fund grwing in clse prximity t ne anther (all ccur, fr example, in stands f frests f a few hectares area n the Prt Hills, western Banks Peninsula, near Christchurch) each has rather different germinatin behaviur. The differences may relate as much t phylgenetic histry as t eclgy, althugh the seed dispersal mde f each species has sme bearing n the germinatin behaviur. In the ensuing discussin seeds in the standard treatment are being cnsidered, unless therwise indicated. Ripgnum is very interesting in having a capability fr viviparus seed germinatin. The viviparus habit may have develped as an imperative fr early spruting f newly-dispersed seeds that have nly a thin testa and fr which drying-ut is fatal (cf. Macmillan 172). Ripgnum seedlings can establish under dense frest canpies. Large-seededness and vivipary culd be beneficial in these cnditins. The large seeds prvide nutriment t the emerged seedlings fr many mnths, as can be seen when the germinated seeds are kept n sil in glasshuse cnditins. Misture must be maintained arund the seed befre and after germinatin. In mst f the viviparusly-germinated Ripgnum seeds which were bserved the emerging seedlings appeared simply as "green windws" in the testa. Presumably it wuld be hazardus fr germinatin t prceed much further than this in seeds which are bird-dispersed, therwise there wuld be a strng risk that the seedling culd be damaged in the digestive tract f the dispersing bird. Ripgnum seeds left surrunded by the pericarp in the in-fruit treatment did nt prceed past this green windw phase. Pericarps f mst seeds in this treatment remained red and undecayed fr six mnths. The sets f Clematis and Tetrapathaea seeds germinated ver a relatively shrt perid in late winter- early spring. The cause f the germinatin delay at the utset is uncertain. It culd arise frm immaturity f embrys, vercme simply by the passage f time. Further experimentatin is needed t examine specific causes f germinatin delay in these tw species. Within the sets f Muehlenbeckia seeds, again after a delay with an unknwn cause, seeds germinated at a gradual rate until mid May, but nne did s during the cldest part f the winter. Thus, germinatin in this species is affected by a degree f envirnmental cnstraint. Lng drawn-ut, spradic germinatin, extending ver tw and three years, respectively, is evident amng individuals in the sets f seeds f Rubus and Parsnsia. These patterns, with episdes f germinatin in each year and gaps between, indicate that a prprtin f the seeds in each set are truly drmant. That is, germinatin is blcked by smething inherent in the bichemistry f the seeds, which must be vercme by particular envirnmental cues. Furthermre, in a species like Parsnsia, in which sme seeds germinated in the warmer parts f three successive years, the causative factrs f delay (drmancy) are impsed t differing degrees within the set f seeds. Successive cld perids in three winters appear t be necessary t vercme the deepest drmancy. After a delay f abut tw mnths at the utset, individual Rubus seeds germinated at intervals thrugh the winter; there is n clear indicatin that lw temperature is a requirement fr vercming drmancy. Hwever sme chilling experiments need t be dne t ascertain the rle f temperature in the germinatin f this species. Bth Rubus and Parsnsia presumably culd maintain seeds in sil seed banks fr mre than a year. If their seeds withstand dry cnditins, they may have lnger-term, persistent seed banks. The hard-cated seeds f Calystegia culd survive, dry r mist, in sil seed banks fr at least five years. The very slw trickle f germinatin recrded fr seeds f this species shws that, althugh mst seeds are impervius t water, this impenetrability is vercme frm time t time by natural prcesses. In the cntext f the present experiments micrbial decay f the cat seems t be the means by which the physical cnstraint f the seed cat is vercme. In nature the cats culd be abraded in streams, r decayed micrbially.

10 Burrwserminatin behaviur f seven NZ vines 101 Dwnladed by [ ] at 11:41 01 December 2017 Eclgy in relatin t seed prperties Ripgnum appears be the mst shade tlerant f the seven vine species cnsidered here. Hwever, seedlings f each can establish beneath frest canpies and seedlings f Clematis, Parsnsia, Muehlenbeckia, and Rubus can be very abundant n the grund in Banks Peninsula frests (C. J. Burrws unpubl, data). These fur species and Tetrapathaea, as adults, are very cmmn frest margin species, as well as ccurring in frest canpies r scrambling ver trees adjacent t frest gaps, s that they may be thught f as requiring well-lit habitats. It is true that their seedlings respnd well by grwing rapidly in well-lit cnditins but they are als, in the seedling stage, mderately shade tlerant. Clematis and Tetrapathaea are lwland species requiring warmth and tlerating nly light frst. Tetrapathaea is the least tlerant f lw temperature. Parsnsia is als cnfined t lwlands; it is slightly mre cld-tlerant than Clematis. Muehlenbeckia and Rubus bth extend much further inland and int the uplands than any f the species just nted. Their germinatin behaviur, in relatin t winter cnditins, accrds well with this. Ripgnum, althugh its seeds, t, germinated ver the Christchurch winter, is cnfined t lwland habitats and tlerates nly light frst. As well as being fund in shaded frest cnditins it is mst abundant in mister parts f New Zealand (in cntrast with the ther species studied, except, perhaps, Rubus). The seed prperties and germinatin behaviur f Ripgnum accrd well with the habitat preferences f the species. It shuld be nted that large-seededness des nt cnfer rapidity f grwth n the yung seedling f Ripgnum. Like Hedycarya arbrea, a relatively large-seeded tree species als able t establish in deep shade (Burrws 15c), the seed and seedling grwth prperties f Ripgnum appear t be crrelated with persistence, in cnditins where seedlings f light-demanding species wuld be strngly disadvantaged. As Leishman & Westby (14) shw, the ability t develp tall, etilated stems, when ther light-demanding species cannt, is a particularly imprtant attribute f large-seeded, shade-tlerant seedlings. Calystegia adult plants are winter-deciduus in inland lcalities where relatively severe frsts are experienced. Its seed characteristics and germinatin behaviur seem t be related t survival f seeds in peridically dry places, mre than ability t withstand cld, althugh they are, when dry, als cldresistant. Hwever, Calystegia des nt extend quite as far inland, r t such high altitudes, as Muehlenbeckia and Rubus. Althugh the ptential frubus, Parsnsia, and Calystegia fr frming quite lng-term seed banks seems apparent, the results suggest that this may be at the expense f verall seed viability within a chrt (many seeds die as they lie, waiting t germinate). It is pssible that thers amng the vine species studied here culd als persist in sil seed banks if the seeds were buried r dried. Experiments n Clematis, Tetrapathaea, and Muehlenbeckia seeds are needed t test this. Ripgnum seeds have germinated when buried 5 cm deep in sil (C. J. Burrws unpubl, data), s it is dubtful that it culd frm sil seed banks f mre than a few mnths' duratin. Its seeds can persist, within fruit n the parent plant, fr up t eight mnths, hwever. Persistent seed banks n the parents (fr tw mnths r mre) als ccur fr Calystegia, Clematis, Muehlenbeckia, and Tetrapathaea. Further wrk is needed t establish their duratin. ACKNOWLEDGMENTS Funding frm the Fundatin fr Research, Science and Technlgy and the University f Canterbury is gratefully acknwledged. I am als grateful t technical staff f the Department f Plant and Micrbial Sciences, University f Canterbury, fr helping t maintain glasshuse facilities, t my wife, Vivienne, fr assistance with glasshuse recrdings, and t Nancy Gh fr her skilful typing. REFERENCES Allan, H. H. 161: Flra f New Zealand. Vl I. Wellingtn, Gvernment Printer. Burrws, C. J. 18: Patterns f delayed germinatin in seeds. New Zealand natural sciences 16:13-1. Burrws, C. J. 13: Vivipary and effects f maternal tissues n germinatin in sme New Zealand seeds. Canterbury Btanical Sciety jurnal 27: Burrws, C. J. 14: The seeds always knw best. New Zealand jurnal f btany 32: Burrws, C. J. 15a: Germinatin behaviur f the seeds f the New Zealand species Fuchsia excrticata, Griselinia littral&, Macrpiper excelsum, and Melicytus ramiflrus. New Zealand jurnal f btany 33:

11 Dwnladed by [ ] at 11:41 01 December New Zealand Jurnal f Btany, 16, Vl. 34 Burrws, C. J. 15b: Germinatin behaviur f the seeds f the New Zealand species Aristtelia serrata, Cprsma rbusta, Crdyline australis, Myrtus bcrdata, and ScheJTera digitata. New Zealand jurnal f btany 33: Burrws, C. J. 15c: Germinatin behaviur f the seeds f six New Zealand wdy plant species. Zealand jurnal f btany 33: Cartman, J. 183: The cultivatin f Clematis spp. Unpublished thesis fr Natinal Diplma f Hrticulture, Lincln University, New Zealand. Dawsn, J. 188: Frest vines t snw tusscks. Wellingtn, Victria University Press. Leishman, W.; Westby, M. 14: The rle f large seed size in shaded cnditins: experimental evidence. Functinal eclkv 8: Macmillan, B. H. 172: Bilgical flra f New Zealand 7. Ripgnum scandens J.R. et G.Frst (Smilacaceae) supplejack, karea. New Zealand jurnal f btany 10: Mre, L. B.; Edgar, E. 170: Flra f New Zealand. Vl. II. Wellingtn, Gvernment Printer.