AN AGRICULTURAL EXPERIMENT IN THE UNPROTECTED SALT MARSH Part II. S. Bottema, T. C. van Hoorn*, H. Woldring and W. H. E.

Transcription

1 AN AGRCULTURAL EXPERMENT N THE UNPROTECTED SALT MARSH Part S. Bottema, T. C. van Hoorn*, H. Woldring and W. H. E. Gremmen CONTENTS r. NTRODUCTON z. WEATHER CONDTONS AND TDAL FLOODS 3. PROGRESS OF THE RESEARCH The Season 3.2. The 1977 Season 3 + The Season 4. GROWTH AND YlELD OF THE CROP 5 GROWTH CURVES 6. GENERAL REivCARKS 7. REFERENCES * Leensterweg 24, EC Ulrum, Netherlands.

2 S. BOTTEMA, T. C. VAN HOORN, H. WOLDRNG &W. H. E. GREMMEN l. NTRODUCTON The possibility of growing rops on brakish soil in onnetion with habitation in the Duth oastal (te1pm) area is disussed by van Zeist et al. ( l 976). That study was the result of experiments arried out for seven years on the unproteted salt marsh north of the Westpolder (muniipality of Ulrum provine of Groningen). The experiments wer ontinued for another three years ( l ). For detailed information on the aim of this study, the loation of the experimental field (fig. 1) and the desription of the agriultural rops used, the reader is referred to the report of the preeding seven years (van Zeist et al., 1976). The experiment generally is a ontinuation of the field study mentioned above. The loation and dimensions of the plots are the same. A sheme of the field is given in fig. 2. Apart from attempting to onf1rm the preeding results some new rops were tned. Some aspets were now studied in more detail. n addition to the rops grown during the period l 969- l 97 5, Tritiw111 aestivt1111, T. spelta, T. dio11111 and Brassia rapa were tried. Another ruiferous plant sown was Brassia oleraea. Seed was olleted from plants originating from Bretagne. Attention was paid to differenes in white-flowered and (oldfashioned) blue-flowered flax, whereas also tworow and four-row barley were ompared. Ca111eli11a sativa was no longer inluded in the experiment as the behaviour of this plant had already proved very onstant under the given onditions. For some rops growth urves are given (figs. 4-8). Harvest data are also given in these figures. For those rops for whih no growth urve is presented harvest data are mentioned in the text. General results are presented in table WEATHER CONDTONS AND TDAL FLOODS Preipitation, minimum night- and maximum daytemperatures are shown in fig. 3. These data were kindly supplied by ng. K. P. Wouda (Mansholt's Veredelingsbedrijf, Ulrum). High tide exeeding 2.05 m was reported by Rijkswaterstaat, diretie Friesland (see table r ). Fig. 1. Loation of the experimenal field (indiated by an arrow) in the undiked marsh near the Westpolder. SOm 128

3 An agriult11ral experi111e11t i11 the 1111proteted salt 111arsh TABLE 1 High lides in metres above 2.05 m NAP, reorded at Lauwersoog date height date height date height Li Sitatissi (A-5 ; B-5) (B-7) The white-flowered variety (A-5 ; B- 5) (fig. 4) and the blue-flowered flax (B-7) started germinating in the first week of May. By mid-j\fay 25 % had germinated on A-5 and on B-5 and B-7 50% and 40% respetively. After a rainy period in June another 5 % flax-seed germinated, so that plants of m ourred with plants of 3-5 m in between. From the last week of June until the harvest on July zznd the flax did not grow visibly. The blue-flowered variety was a. ro m shorter than the white variety and also had a lower seed prodution. Viia Java var. 111i11or (A-r; B-r) n field A-r not a single bean sprouted. n B-1 Jess than ro% germinated but only a few plants remained, reahing about 5-ro m towards the end of June. By mid-j uly they were up to r 6 m tal!. Some produed poor flowers together with yellow leaves.,-\ week later all plants were gone. 3. PROGRESS OF THE RESEARCH 3. r. The r 976 Season Ge11eral The very dry weather strongly influened the growth of the various rops. t was observed that this drought effet was not of equal intensity all over the field. t was most pronouned in the northwestern orner and least in the southeast. Various fators would have been responsible for this gradient. From north to south there is a drainage effet, as the steep erosion edge was not more than about 3-5 m away from the north side of the field. n the ourse of ten-year experiments the edge slowly rept towards the field as the winter storms aused it to ave in. Northwestern winds, that are quite ommon in the area, bring salt as they blow over the Wadden Sea. This has a marked effet on the front row of plants whih have a protetive funtion for the plants behind. The resulting salt ontent is also higher in the northern part beause of the drainage effet. Thus onditions an vary quite a lot on plots of a few square metres. Crops were sown on April 9th, apart from Tritit1111 spelta and Brassia rapa, that were sown on April r 5 th. Horde11111 distih11111 (B-8) Four rows of two-row barley were sown on plot B- 8 to ompare with Hordet1111 1mlgare. At the beginning of May this barley was up to ro m high. Two months later these plants measured about 40 m and were produing ears. The development of the ears however was very poor, giving hardly any yield. The sparse yield was finally eaten by birds. Horde11111 vulgare tetrastih11111 (A-4; B-4) Four-row barley was sown on A-4 and B-4; only z 5 o gr. of the breed Bigo was available. Germination was good but in the ourse of May some plants dropped out. The plants in the B-field were about 10 m taller than those in the northern (A) plot. This differene is refleted to some extent in the yield. Ave11a sativa (A-3; B-3) Germination and initial growth were better on the northern plot. n the ourse of June, however, the rop on the southern (B) plot inreased and surpassed that on the A-plot, being about ro% higher. This is also visible in the very moderate yield, har vested on July zznd. Pa11i iliae11111 (B-6) Germination took seven weeks and then only

4 .. S. BOTTEMA, T. C. VAN HOORN, H. WOLDRNG&W. H. E. GREMMEN EXPERMENTAL FJELD :; ' E A-1 A-2 A- 3 A-4 A- 5 A-6 A-7 A- 8 :i ' E :i B-1 B-2 B-3 B-4 B-5 B-6 B-7 B-8 unaltered grassland 0 1m TABLE 2 Data on input and yield on the experimental field in the undiked salt marsh near the Westpolder. 0 'O :;:; :; Q) :;;, 'O Ol Ol a. <l Ol Ol ro.i:: '.. ::J... :; 'O 'O u a. Q) Q) (ij :;;, :;;, u Ol ::J 'O a. Q) :;;, 0... :; a. ' 'O Q) :;;, 'O Q) :;;, 'O <l ro.i:: ::J u (ij u Ol A 1976 B Linum usitatissimum fl. white Linum usitatissimum fl. blue Viia faba var. miner Hordeum distihum 65 Hordeum vulgare tetrastihum Avena sativa Panium miliaeum Tritium aestivum Tritium spelta 40 Tritium dioum Brassia rapa var. silvestris

5 AJJ agri11t11ral experiment ijj the 1111proteted salt JJJarsh... Fig. z. Experimental field. amounted to about 5 % Some plants reahed about 45 m. By the seond week of August the leaves had turned brown. On September 3fd a very poor rop was harvested. Triti11111 aestiv1m1 (A-2; B-2) Germination and growth was very poor. At first development on the northern plot was better, but this rop was overtaken in terms of growth towards the end of June by the rop on the B-plot. Tritiuv1 spelta (A-7) 40 gr. were sown; 10 plants developed and reahed 40 m., All the spelt plants died before produing any ears. Triti11111 diom1111 (A-7) Germination was about 30% and the biggest plants reahed m. The ripening proess took more time than in the other ereals. Brassia rapa (A-8) The summer rape obtained from the Zentralinstitut for Genetik und Kulturpflanzenforshung, Gatersleben (D.D.R.) was named there: Brassia rapa L. var. silvestris (Lam. Briggs) f. preaox (DC)Mansf listed under assortments number BRA 4074 and alled Maleksberger. The amount of seed was suffiient for half a plot. The first seeds germinated at the beginning of May. By mid-june the plants were m high. A month later maximum size was 75 m and part of the rop was already ripe. Brassia oleraea (A-8 ; B-8) Cabbage was sown in two rows. Germination was good. The plant is biennial and thus flowers after one year. Some of the plants were left in the experimental field, some were taken by one of us (T. C. van Hoorn) and put in a garden to allow for any neessary restoking The 1977 Season Li11t1JJ1 11sitatissi111un1 (A-6; B-6) (A-8) White-flowered flax (A-6; B-6) and blue-flowered flax (A-8) germinated very well. There was a lear differene between the white flax on plot A-6 and that on B-6, while the blue flax lagged behind the white variety on the orresponding plot. n the 0 'O 0 'O Qi Qi ;;, ;;, Ol :; 'O :; 'O Ol 0. Q) Ol Ol 0. Q) ::: ' ::: :::. ::: ::: :.::: ::: :::J... :.::: :; 'O :::J Ol (,) :; 'O (,) 0..li:: Qi Q) 0. (il Qi Q) ::: (il ;;, ;;, (,) ::: ::: ;;, (,) ;;, '.... Ol.li:: ::: 0 :!2 0 :!2 Q) Q) ;;, ;;, :; 'O :; 'O Ol Ol 0. Q) Ol Ol 0. Q) ::: ' ::: ' :::. ::: ::: ::: :.:::. :::J... :.:::... :; :!2 'O Ol :::J (,) :; 'O :!2 Ol.li:: (,) 0. Q) Qi (il 0..li:: Qi ::: ;;, ::: (,) ::: ;;, ;;, Q) (il (,) ::: ;;, A B A B

6 S. BOTTEMA, T. C. VAN HOORN, H. WOLDRNG & W. H. E. GREMMEN ourse of the season, however, the blue variety developed a better average size but the white flax suffered from Thrips. Ripening took longer for the blue variety than for the white one. Viia faba var. 1J1inor (A-2; B-2) Germination was very good. The development is shown in fig. 5. The beans were harvested on September r 5 th. Horde11111 distih11111 (A-r) Two-row barley was sown on the plot where onditions are most unfavourable as regards sea-wind exposure. Half the grains sprouted. Some plants developed well, attaining 85 m maximally. The yield, however, was very low. Hordet1111 1;1tgare tetrastih11111 (A-5 ; B-5) This barley developed well, although damage was inflited by birds and hares. There was no visible differene between the northern and the southern plot. As gei:mination took plae at two suessive times, ripening also showed two groups. Ave11a sativa (A-4; B-4) Development on both plots ran parallel. The yield from the northern plot was a little better. Harvesting took plae on August 24th. Pa11i iiae11111 (B-7) Milled did not germinate. Triti11111 aestiv11111 ( A-3; B-3) Germination was not so good as in four-row barley and oats. At first growth was better than in the other ereals. Later growth delined, ears were formed with diffiulty and leaves turned yellow at the tip. The rop was harvested during the last week of August. Brassia rapa (B-r) Rape seed grew very well and was harvested on July 25th. Brassia oleraea (A-8; B-8) The abbage, sown in 1976, had developed ro plants. By mid-may they were about 8-ro m high and a week la ter three of them flowered. At the end 132 of June they measured 30 m and fruit-setting had started. Flowering la sted up to J uly r 1 th. The seed prodution was 20 gr. The yieldinput ratio is not known, beause the input was not weighed. The yield of a Brassia oleraea plant originating from Helgoland and growing on fertile sandy soil, inland, amounted to 94 gr.! 3.3. The 1978 Season Li sitatissi (white-flowered A-7; B-7) (blue-flowered A-7; B-7) n the plot B germination was more even. n the A plot growth was not so good and the blue flax still remained shorter than the white variety. Ripening and harvesting differed by one week : August r 5 th and 2 r st for the white and blue varieties respetively. Viia faba var. 1J1i11or (A-3; B-3) The northern plot lagged behind from the beginning. At an average length of r 5 and 2 5 m respetively plants on the two plots showed flower buds in the third week of June. Towards the end of Jul y plant-lie (aphids) appeared in the northern part. The biggest plant on the B-plot measured r r 2 m. Harvesting took plae on September 2 r st. Hordet1111 distih11111 (A-8) Two-row barley sprouted regularly. During the first few days of June the leaves started to turn yellow and growth stagnated. The few ears that developed stayed in the enveloping leaves. The remaining ears that bore seed were eaten by animals. Horde1 11 v11gare tetrastih11111 (A-6 ; B-6) Four-row barley developed well but suffered, as usual, from birds eating the seeds and hares eating the tops of the young plants. At the beginning of June some of the plants turned yellow. The ripening rop was partly eaten up by birds. Part of the rop was harvested on August 2 rth and the rest on September 9th. Ave11a sativa (A-5; B-5) Oats developed well, espeially on the southern plot. J\faximum length was 90 m. Befare the harvest, birds took most of the rop. The data in table z are therefore rather meaningless.

7 O M.l.RtK 10 mr:i! ru an 11 mnl Hllt 5.S \ (11.1an HH] M.1.Y ls [mro 19m'J JVH H 1:1n \ml an s sr.111 ULY ll.s 1:u1 [mun H ir.'llf AU&US ZO,S r11n (l!ltaft 82 ll\ø )., L, J L 1._.9._7 6.._, JUi! 1m UUS O HUCK ll 11 (iraft l nr.ri l J.Pl!L '1,S i:u1 (111u 45 mm ) HU SS (r.un 491lnl JU lur.1111 (1111an Sl1M1) llly ns r.in lr:11an 1>1M1 Uit S nm tni1a1 821m1J MAU JUi( Jill.,..UUS -12' L J J L l,_,9-7,, 7 10 E ELO APRL JO mm \r.'.uft ) MAY SS mm [11111n lhtnj JJNE 55 mm (1111u 5lM JULY 48.S mm tmun :1) AUGUST 53,Smm (l!uu l mnij JO JO MARCH APRL MAV JUlfl JULY AUGUS Fig. 3. Preipitation and minimum and maximum temperatures from Marh 1st to At:1gust 1st, 1976, 1977 and 1978, reorded at Mansholt's Veredelingsbedrijf (Plantbreeding firm) in the \Xlestpolder. The mean temperatures and preipitation (for the per iod ) are af ter Atlas van Nederland, Plates V-1 and V

8 S. BOTTEMA, T. C. VAN HOORN, H. WOLDRNG &W. H. E. GREMMEN Pa11im111 liliae11111 (B-8) Millet failed again although some plants reahed m. There was no yield. Triti11111 aestivt1111 (A-4; B-4) Wheat on the B-plot sprouted rather well. By mid May, however, many plants had <lied. n the seond week of July birds had eaten the remaining rop. Brassia rapa (B-2) Summer rape was a suessful rop, flowering abundantly at the beginning of June. Damage was inflited by aterpillars of a Pieris speies, and also the pods were peked out. Part of the rop was harvested at the beginning of August, the remaining part later. Brassia oleraea (A-2) Germination was a little irregular, espeially on the seaward side. Plants reahed about 20 m maximally. As this plant is biennial, no prodution is to be expeted this year. 4. GROWTH AND YELD OF THE CROP The information obtained from the seasons r will be disussed and ompared with the results of the preeding seven years. The data are treated the same way as in van Zeist et al. ( l 976). n table 2 the data for the various rops are shown, apart from Brassia oleraea. Plotted are the input in grammes, the yield in grammes, the yieldinput ratio and the alulated yield in kgha. As previously the northern (A) plots were lightly manured with dried attle dung. Any positive effets of manuring was generally overshadowed, however, by the deleterious effets of salt northwestern winds. Li sitatissi1J From 1975 onward, two flax varieties were used: the white-flowered variety that is grown nowadays and a blue-flowered variety that was grown previously in the Netherlands but that is no longer ultivated at the moment. The 1976 report states There is no signifiant differene in yield apparent between the white- and blue-flowered flax. This is onfirmed by the results of the three following seasons. n 1976 and 1978 the seed prodution of blue flax was lower than that of the white flax whereas in 1977 it was higher (table 2). The yields also fall within the limits of the first investigations. The growth-urves (fig. 4) show that the white flax is on average about 10 m higher than the blue variety. On sandy soil in Orvelte (prov. Drenthe) the two rops also demonstrate the same tendeny. Blue flax also develops more slowly than white flax and ripens later. Via f aba var. minor The Celti bean had two suessful seasons as against one season of omplete failure. An unsuessful season is aused by unfavorable onditions during the germinating stage. When flooding with salt water or a very dry spring (inreasingly high relative salt ontent) our during and after sowing then generally no plant will appear. When enough preipitation falls then even aphids annot ruin the rop and a reasonable to suessful yield is ensured. Fig. 5 demonstrates the differene in growth between a failing rop and two suessful rops. The manured plots did not show a better yield, whih is explained by the negative influene of the sea wind from whih the southern plots are sheltered by the :rops on the north side. The A and B urves, dra\v n in fig. 5, learly demonstrate this effet. Aording to Kreuz (Aerboe et al l 9 30) the original Mansholt Celti bean averaged o.8 gr. and had a yellow-brown olour. Offspring from Celti beans from the same farm that were grown on the marsh averaged o. 3 5 gr. in l 978 and were brown in olour. From this strain, grown on poor sand, inland, the beans averaged o. p gr., whereas the olour was yellow. Sandy soil, heavily manured with attle dung, gave an average of o. 37 gr. t seems that the olour is dependent upon the soil but also upon the time of harvesting (ripening). Horde11111 distih11111 a11d Horde11111 vulgare tetrastihum The results of the experiments during the period l suggested that four-row barley was better adapted to brakish onditions than two-row barley. This is onfirmed by the results from the three last seasons. From both barleys were grown and ompared (fig. 6). The piture is often obsured by

9 L NUM USTATSSMUM white + blue flowered variety 1E u BO plots A and B ha rvest _ white ( 8 ) eblue (8) ' t h't,, -,-:-..= W e.... 1,...,,.: wh1te (A) ; W h't e e blue,, i 1 blue 11 V, may june july august Fig. 4. Growth urves for Li sitatissi))lJ1, white and blue flowered variety. 1E u.. :7> '..:. 120 VCA FABA var. MNOR B 159 plots A and ,' ,,, A 159,' BO 197B ;,;.i' t apri dead,, ', ',, ',,, ' ; -- ; ;::; _,,,.,,,,,.,,,.,.,, ;'::.:.:..: -:::.:.: :...,_ t B,,.,-_,,,..,,,,. '.;,.; A 219 A may june july august i Fig. 5. Growth urves for Viia Jaba var. ))i11or

10 1 E '-'.!: -. = Cl> BO HORDEUM VULGARE, H. DSTCHUM harvest t de ad,, _.....-f dist evulg evul g edist. t d' 1st. 20 _-:: : april may june july august Fig. 6. Growth urves for Horde11111 vulgare and H. dislihu111. TRTCUM AESTVUM E '-' :; harvest X eaten by hirds,.,,,.,,,,,,,...,,, X -. = Cl> may june july august 136 Fig. 7. Growth urves for Trilimm aesli1j11111.

11 An agriult11ral experilllent i11 the 1111proteted salt 111arsh BRASSCA RAPA BO r E u ::: harvest ;.;.;.;.;, ; ; may june july august Fig. 8. Growth urves for Brassia rapa. damage inflited by birds eating barley during sowing ånd harvesting. Nevertheless, the fats point onviningly to Horde11111 v11gare tetrastih11111 having advantages over H. distih11v1. n three seasons four-row barley delivered three harvests as against one of two-row barley, and even when both types of barley produed a harvest the yield of the four-row barley was seven times as great as that of the two-row barley. Avma sativa The behaviour of oats during the period under disussion is the same as during the first study (van Zeist et al., 1976). Oats sometimes give a good yield but on fewer oasions than is the ase with fourrow barley. Good examples are the 1975 season (see first study) and the l 977 season. Oats will produe a good yield when the preipitation is high and temperatures are not too low. Rainfall must be of the order of 50 mm or higher for April as well as for May under the Westpolder onditions. Pa11i iliae11111 Millet was sown in l 97 3 for the first time in the experimental field. Beause it was not known whih variety of millet was best adapted to Northwestern European onditions and to the Westpolder espeially, several varieties of different origin were sown. Of these varieties La Plata did better than the others. n 6 years there were only two harvests, eah with a limited yield : in 1975 the output was 5.5 times the input, and in l 976 some seed, a guarter of the input, was harvested. n the other years hardly any millet plants developed. Pa11i iliae11111 was found by van Zeist in non-brakish surroundings (1974) in Vlaardingen Broekpolder in a house site dating from the ron Age (3 70 ± 70 B.C.). The experiments show that in brakish surroundings millet annot be grown. Triti11111 aestiv11111 Bread wheat was inluded in the program from 1973 onward. On some oasions a few plants developed produing a yield a few times greater than the input (1975, 1977). When the other ereals produed high yields the bread-wheat yield remained at 10-20% of the Avena and Horde11111 v11!gare prodution. For growth urves see fig. 7. The results demonstrate that bread wheat annot have been grown on the unproteted salt marsh under onditions omparable with those of the Westpolder. The many oastal sites investigated by van Zeist (1974) do not show any bread wheat during prehistori

12 S. BOTTEMA, T. C. VAN HOORN, H. WOLDRNG &W. H. E. GREMMEN 13 8 times, apart from a find from Ouddorp on Goeree that ould well have been imported. Triti11111 spelta Spelt did not give any results. There are no indiations that spelt was grown in the oastal area. Tritim111 diom11 Emmer wheat grown for three suessive years only produed same plants in 1976 but the yield was very low, about half the input. Van Zeist mentions same emmer wheat for prehistori Paddepoel (1974) and for pre-roman ron Age Middelstum (verb. omm.). On the unproteted salt marsh emmer wheat would not have been suessful. Brassia rapa Rapeseed (summer rape, Brassia rapa = Brassia ajjjpestris forma a111ma aording to Bengtsson, von Hosten & Loof, 1972) proved a very satisfatory rop for the unproteted marsh. Growth urves are very similar to those of CaJJJeli11a sativa (van Zeist et al 1976). Summer rape has a reasonable yield even under unfavourable onditions. Van Zeist mentions Brassia ajjjpestris (rapa) for 8 out of 10 prehistori 0 _s in the Duth oastal area. The relatively short vegetation period of this plant and its immunity to ertain hanges in loal onditions make it a very reliable rop (fig. 8). 5. GROWTH CURVES n the first artile on this su bjet (van Zeist et al 1976) growth urves were shown for CaJJJeli11a sativa (not inluded in the last three years of the experiment), Li sitatissiJJJt1111, Avena sativa and Horde n the growth urves the average height of the rop was plotted against time. t must be stressed that suh a urve does not neessarily give any information about the yield. Even if a few seeds germinate on a plot, it is still possible to draw a growth urve. f these plants die before the harvest there is no yield at all, but there is still a growth urve. On the other hand the growth urve does say something about the state of health of the rop and this state is aften also an indiation of the produtivity. The urves (figs. 4-8) are based on information obtained during the last three years and are shown for three purposes. One series inluded urves not presented in the first study, viz. Viia Jaba var. JJJi11or (fig. 5 ), Brassia rapa (fig. 8) and TritittJJJ aestivt1111 (fig. 7). Seondly, urves are shown whih ompare the growth behaviour of two different types of the same rop, viz. white and blue varieties of flax and the two-row and four-row breeds of barley. Finally the differene between the A and B plots is demonstrated. The growth urves of blue- and white-flowered flax run parallel, there being in general a small advantage for the white-flowered variety (fig. 4). The growth urves for the seasons l ompare well with those of the preeding years. Van Zeist et al. ( l 976) point out the similarity of the behaviour of Viia Jaba var. t11i11or and oats and barley: either a steeply rising urve during June and July or a ompletely flat urve, running horizontal for both months. This learly demonstrates the differene in prodution, these two urves indiating either a good rop or a omplete failure (due to germination problems) aordingly. n fig. 5 the growth urves for Viia Jaba var. 111i11or are shown for the three seasons and moreover for both plots, the A-plot being doser to the sea and the B-plot sheltered by the A-plot. Espeially when the plot has a low number (fig. 2) it is one of those most exposed to sea wind. With the aid of these separate growth urves an attempt has been made to indiate the sensitivity of a rop to the rniro-habitat of a sandy ridge in the salt marsh. Not every square metre of the arable area is equally suitable for optimum rop performane. The differenes in development and yield of two-row and four-row barley have already been disussed in this study. nformation on the absolute size an be derived from the growth urves. The Horde11111 v11lgare urve rises muh more steeply and at an earlier point in time than the HordettJJ distih11jjj urve (fig. 6). Not only do the plants develop more rapidly, they also attain a muh greater height. Bread wheat has a growth urve (fig. 7) whih resembles that of two-row barley. Curves do not rise steeply but tend to run rather smoothly. Bread wheat is also omparable with two-row barley in terms of prodution or lak of yield.

13 A11 agrimlt11ral experi111e11t i11 the 1111proteted salt 111arsh The growth urve of Brassia rapa (fig. 8) shows the same harater as the other ruiferous oil plant Ca111eli11a sali1ja. The growth urves run parallel with a steep urve in June demonstrating a relatively short growing season. The similarity in the growth urves demonstrates that summer rape 1s not easily influened by external fators. 6. GENERAL REi\iARKS The experiments arried out during the last three years giv.e us supplementary informa_tion on the possibility of raising rops on unproteted salt marsh during the prehistori and early historial period rather broadly desribed as the tc1pe11 period. n the results were ompared with the results from Korber-Grohne's experiments at Cappelersiel ( r 976, pp r ). Kbrber-Grohne onluded that Ca111eli11a sati1ja was the rop most resistant to salt and wind, followed by Horde mlgare, Viia faba var. 1J1i11or and Lil sitatissi Van Zeist el al. share her opinion on Ca111eli11a but think that Horde11111 is more sensitive than Li n the first part of the experiment f-lorde11111 distih11111 was used twie instead of Horde J11gare, as seed of summer-sown f-lorde J11gare was sometimes diffiult to obtain (at the moment four-row summer-barley is not grown on an eonomi sale in the Nethrlands). This resulted in onfusion of the yield weights and the growth urves. After planting four-row barley for another three years ( r ) it seemed to be a more reliable rop than was thought at first. Van Zeist (1974) identified finds of barley from prehistori times in the Duth oastal area as belonging to f-lorde J11gare. This faet together with the results from the agriultural experiment make it lear that Horde11111 distih111;1 is not so well adapted to the brakish surroundings of the oastal lowlands. During the te1pen period four-row barley would have guaranteed a moderate to reasonable yield. From table z it an be seen that Li does not show so muh variation in produtivity as Viia Jaba var. 111i11or, but that the Celti bean an have a very high prodution as well as omplete failures. t should therefore still have been an eonomially important rop at that time. Considering all the problems that a prehistori le1pe11 farmer ould enounter, the moment at whih a problem arose was very ruial. Thus a flood in Jul y would have damaged flax more than the Celti bean. Had the flood taken plae in April during or just before sprouting then possibly neither flax nor beans would have survived. \Xhen we onsider the rops used in the experiment we an state that under ertain onditions Horde mlgare, A1Je11a fat11a, Ca1J1eli11a sati1ja, Brassia a111peslris, Viia Jaba var. 1J1i11or and Li sitatissi were grown or ould have been grown by farmers who inhabited the unproteted salt marshes along the oast. Crops like Horde11111 distih11111, Trili11111 aesti1;11111, Tritim111 spelta, Tritim111 dio11111 and Pa11i iliae11111 were guite likely less suitable or unfit for brakish onditions. f they were grown in the unproteted area it must have been guite far from where the influene of the sea prevailed. The experimental fi eld at Westpolder is situated where onditions are rather extreme. One must assurne that the!e1pe11 farmers met with fewer diffiulries than we did. t is nevertheless remarkable that within a distane of a few metres differenes within one rop are obvious. This must partly be due to exposure to the northwestern winds oming from over the sea, and partly also beause of differenes in omposition of the ridge. The northern part of the ridge is situated losest to the sea and there most of the sand is deposited. Further from the sea the soil ontains more lay. The rops do best on a mixture of sand and lay. The purer the sand the more severe are dry periods whih also ause an inrease in relative salt ontent. On pure lay (see the results of suh a field, desribed in van Zeist et al 1976) the salt brought in by the sea during the winter is not easily washed out in the spring. Fields suitable for farming were relatively sare during the te1pe11 period as the farmers ould only use part of the higher ridges along the oast and alongside streams. On the other hand the original kwelder (marsh) area was several kilometres wide, with the sea at some distane. Thus farming onditions were not too unfavourable.

14 S. BOTTEA, T. C. VAN HOORN, H. \V OLDRTNG & W. H. E. GREMMEN 7. REFERENCES AERBOE, F., J HANSEN & TH. ROEER, Ha11dlmrh drr La11d1JJirtshaft,. Berlin. ATLAS VAN NEDERLAND, Samengesteld door de Stihting \Xletenshappelijke Atlas van Nederland. (Atlas of The Netherlands. Compiled by the Foundation for the Sientifi Atlas of The Nerherlands). 's-gravenhage. BENGTSSON, L., A. VON HOFSTEN & B. LOO F, Borany of Rapeseed. n: L. A. Appelquist & R. Ohlson (eds.), Rapeseed. Amsterdam. KO R BER-GROHNE, u Ceobota11ishe U11termdm11ge11 ruif der Fedderse11 Vierde. \Xliesbaden. VAN ZEST, w Palaeoboranial studies of settlement sires in the oastal area of the Netherlands. Palaeohistoria 16, pp VAN ZETST, \V., T. C. VAN HOORN, S. BOTTElA & H. \V OLDRNG, An agriultural experiment in the unproteted salt marsh. Palaeohistoria 18, pp