Veget Hist Archaeobot (1992) 1:43-52 VegetationHistory and Archaeobotany © Springer-Verlag 1992 Palynological evidence of Azolla nilotica Dec. in recent Holocene of the eastern Nile Delta and palaeoenvironment Suzanne A.G. Leroy Laboratoire de Pal~og~ographieet Pal~ontologie,Institut G~ologique, Universit~ Catholique de Louvain,3, place Louis Pasteur, B-1348 Louvain-la-Neuve,Belgium Received November 25, 1990 / Accepted March 25, 1991 Abstract. Megaspores, microspores and massulae of the free-floating fern, Azolla nilotica, were found in Late Holocene sediments obtained by coring in the eastern Nile Delta. Nowadays, the nearest station for this fern is southern Sudan. The determination of the species is based on spiny projections on the megaspore body and on the verrucate microspores. Palynological studies reveal that the habitat of the fern consisted of extensive papyrus marshes, now disappeared. Several causes for the disappearance of the fern from the Nile Delta are proposed amongst which the most probable is human influence which has completely modified the vegetation and the hydrology. Key words: Holocene - Azolla nilotica spores Palynology - Nile Delta Introduction The Azollaceae are free-floating aquatic ferns. They grow in quiet waters and on mud near sluggish river edges, occasionally forming a dense carpet on the water surface. Fossilized reproductive organs such as megaspores, microsporangia and microspores of Azolla are present in Holocene sediments from the eastern Nile Delta. Identification of the fossil material is made with the help of SEM photographs. The fossils come from the upper metres of several cores taken by the Smithsonian Institution around Lake Manzala. The project is directed by Dr D.J. Stanley. CouteUier (CouteUier and Stanley 1987), Thomas and Pimmel (in preparation) provided the sedimentological data. The aim of this paper is to characterize the fossil fern environment and identify the causes of its disappearance hence the reason for the pollen analyses of the sediment around the Azolla-rich levels. The marshes containing the aquatic fern and associated plants were controlled by freshwater floods originating from as far away as the Ethiopian Highlands and, to a lesser extent, equatorial East Africa. The localisation of those marshes in the delta has changed during the Holocene as the distributary channels of the Nile changed course. In the eastern Delta (Fig. 1), the major distfibutaries, from west to east, are the Mendesian branch, the Tanitic branch and the Pelusiac branch. Nowadays, they are silted up. The nearest present distributary is the Damietta branch, west of Lake Manzala. The present day distribution of Azolla does not include Egypt. A. nilotica Dec. is a common fern in Central Africa, its nearest station to Egypt being southern Sudan. In MacLeay (1955a), we see that these ferns "seem to be more or less confined to the headwaters of the Nile system but have been collected so seldom that their distribution is really unknown". He quotes an herbarium sheet annotation: "floating aquatic on White Nile and its tributaries, upstream from Kosti, 350-700 m, common in still water, White Nile, 1840, D'Arnaud (Herbarium Kew)", i.e. about 260 km South of Khartoum (MacLeay 1955b). The plant has a demonstrated agronomic importance in the third world. A z o l l a is capable of assimilating atmospheric nitrogen. The actual agent of fixation is the symbiotic cyanobacterium Anabaena azollae. It is present in dorsal cavities in the fern leaves. It grows successfully in habitats where little or no combined nitrogen is available. In North Vietnam and Thailand the fern is traditionally cultivated in rice paddies and used as a green manure. The presence of Azolla brings 2-4 or more kg nitrogen / ha per day (Perkins et al. 1985; Moore 1969). In some parts of the world it is considered a nuisance and is removed as a weed. Other uses that have been reported include forage, soap, prevention of mosquito breeding, etc. Materials and methods Position and sedimentological description of the levels with Azolla spores. Twenty cores, 20-50 rn long, were taken in an area of 30 by 20 km around Lake Manzala (Fig. 1). The sediment belongs to prodelta mud, delta front clayey-silt, sand 44 31o30 ' 32o00 I BORINGS: + 31o30 '- ' 32~30' PALYNOLOGY O AZOLLA MEGASPORE [] NO O B S E R V A T I O N R A S - E L - BAR 41 -31030 ` O [] / DAMIETTA 4/v "~ LAKE ~N~N;NZAL~'-~ 4 N (~)$2 ~ . • ~*, "--'.,'~ SE 4 SAID Hm30.000w . . . . . #l _ ~ e ~a 31o00 ' s90 5 10 ,0 .~ ,,5 ." [] i,O -31000 , , ' ' ° " °~° .Y" ~ s2 . 15 MILE5 2,°~. i 31030 , ~3" I I 32o00 ' 0 EL- QANTARA r 32o30 ' F i 8. 1. L o c a t i o n o f the cores in the eastern N i l e D e l t a dunes, marsh or lagoon (Coutellier and Stanley 1987). The sand-size fractions of all the Smithsonian Institution samples from 20 cores were searched for fossil Azolla; 9 yielded megaspores (Table 1). Other available data are: the position of those samples in the nine cores, their granulometric analysis, the 14C ages (some unpublished; obtained by the Smithsonian Institution), and the palaeogeography as deduced from sedimentological studies. Palynologieal studies have been carried out on four cores. The samples chosen are in and around the Azolla-rich levels indicated by asterisks in Fig. 2. In the coarse fraction of sample $2-2 (see Table 1), the spore concentration is very high. This level is at 1.80 m depth (B. Thomas, unpublished). It is composed of organic dark brown clay. This facies is 130 cm thick (2.80 to 1.50 m), overlying a 120-cm-thick peat layer and underlying a 50-cm-thick clayey-silty layer. X-ray reveals vertical roots in the Azolla-rich layer indicating temporary emersion after the accumulation of the layer. Herbarium reference material and palynological slides reference collection. The megaspore belongs to Rhizosperma Meyen section because it is 9 floated. It is represented by two species: A. nilotica Decaisne and A. pinnata R. Br. SEM photos were taken of specimens from the U.S. Herbarium at the Smithsonian Institution (Plate 1). The herbarium sheets sampled are: Azolla pinnata R. Br., forma natans, Welwitsch, Iter Angolense, Mart 1857, n38, US 2424619; Azolla pinnata R. Br., Deutsch Ost-Afrika, Reisen in Afrika 1925-1926, A. Peter, n44388, Aug. 1926, ex Museo Botanico Berolinensi, US 1755196; Azolla nilotica Decaisne, Deutsch Ost-Afrika, Reisen in Afrika 1925-1926, A. Peter, n44569, Aug. 1926, ex Museo Botanico Berolinensi, US 1755198. Several slides from the Montpellier, G6ttingen and Louvain-la-Neuve collections were consulted for the microspores of both Azolla species: A. nilotica Dec, Rukwa Lake, Tanganika, Recolt. Goetze IIII, CFM. L.47, Livingstone. Slide nr 916 Sp.; A. africana, Dougia, Tchad, Recolt. Maley. Slide nr 819 Sp. Based on the megaspore, it is A. pinnata; A. nilotica, Zambia, Geneva Herb.; A. pinnata, Madagascar, Geneva Herb.; A. pinnata imbricata, Pi07, cultivated by C. Van Hove, UCL. Palynology. The palynological content of 17 samples has been analysed (Fig. 2, Table 1). The samples chosen are from different facies, for example: the well-developed prodelta mud in samples Sga and Sgb; the prodelta front in S7a; the peat in samples S6a, S6b and $6c; the continental marshes and lagoons in samples S7b, $7c, S7d and S7e and four samples adjacent to the Azolla-rich level in samples S2a, S2b, $2c and S2d. The sediment is treated to extract the pollen and spores as described in Dricot and Leroy (1989), except for core $2 samples for which zinc chloride was used instead of Thoulet solution. A total of 72 taxa have been determined with the help of 45 Table 1. Sedimentological description and 14C ages of the samples with Azolla megaspores Core Depth Seal. Pal. Granulomet~y Palaeogeography 14Cage Megaspores (In) sample sample clay silt umd years B.P. in% 52 1.30 d enntlnental 1.75 52.2 e 84.2 15.4 0.4 eonthumtal 1830570(3) 49% 2.15 52.3 76.3 23.5 0.2 contineatal 2.2% 2.70 b continental 190 a peaty clay 3.80 82.5 19.8 42.3 37.9 peat 38001:90(3) $6 $7 58 S9 2.00 5.00 5.50 5.65 5.80 6.20 1.00 1.50 1.95 2.00 2.70 3.00 3.30 3.40 3.50 3.70 5.60 3.00 3.50 4.10 5.50 10.00 14.80 22.60 $6.8 S6.9 d e b • $7.1 $7.2 65.0 96.0 82.7 32.6 3.1 14.3 2.4 0.9 2.9 20.0 8.8 12.5 6.5 67.5 84.7 86.0 82.0 13.9 10.6 0.1 7.3 77.9 21.6 75.2 23.6 75.2 23.7 0.5 1.2 1.1 29.3 7.8 23.0 36.5 83.9 29.3 • 57.3 S7.4 $7.5 $7.6 $7.7 $7.8 d e b 34.2 8.3 473 a $8.4 S8.5 $8.6 $8.8 d . c 44.1 54.1 1.4 27.0 0.6 91.0 50.2 49.4 0.4 30.6 59.5 9.9 b • ~ontlnental enntinentld continental continental continental continental conline~m] ¢m~atimmml marine deltafi-ont ddtafront pit>delta mud prodelta mud 1.1 0.4 1.2 continental continental eentimmtal 5.50 513.27 36.0 61.4 7.6 delta front/ enntineatal S16 18.50 S16.g 70.2 29.6 0.2 prodeltamud $22 95.2 2.5 3.3 marine $22.5 19104.70 (4) 2.5% 5.1% 3750i60(4) 23405:90(2) 3805±40(6) cont./marine cont./marine cont./marine deltafront 16.2 17.9 2,1.0 2.80 3.50 S9.4 $9.5 $9.7 marine enntinental continental peat pent with sand pezt with sand 82.7 81.7 74.8 S13 2.80 3.30 4.10 $6.2 $6.6 $6.7 * * * * * * * * * 42301 90 (4) 46951115(4) 374Oi150(3) 51405:80(5) * * * * 4820/80(3) * * 3770~90 (6) Hm 9.20 :1;20 4"15 4"65 m ~ i n ~ mouth * 30.000 of Tanltle branch w(1) M ¢ ~ giv~ as a pczezm of sand size fraction;an mm~risk indicatesone or very few (1)Suez Canal offshore; (2) Smithsonian Radiocarbon, Rockville, Maryland (3) Beta Analytic, Miami, Florida; (4) Stanley, Sheng et al., 1988 (5) Stanley, 1988; (6) Foacault et al., 1989 reference slide collections, mainly in the Montpellier Laboratory. Other palaeontological remains were observed (Table 2) such as : diatoms, Zygnemataceae, Pediastrum, Concentricystes, etc. Their determination and the palaeoecological interpretations are based mainly on Van Geel (1978). A minimum of 230 pollen and spores were counted per spectrum. For some samples the concentration in palynomorphs per gram of dry sediment was evaluated. The data are presented in two types of diagrams: 1.Detailed diagrams (Fig. 3), where the percentage of each taxon is represented. The basic sum excludes spores, undetermined and undeterminable grains. 2.Synthetic diagrams (Fig. 4): the taxa are grouped in ecological classes, and their percentage representation is shown. The basic sum excludes the undetermined and undeterminable grains. The Nile Delta plants produce autochthonous pollen and spores. The delta also receives allochthonous grains on one hand airborne from the Mediterranean region, on the other and mainly - brought by the river from the tropical (80% of the discharge) and equatorial regions, by the Blue Nile and the White Nile, respectively. For the detailed diagram all the taxa are ordered in seven groups (Fig. 3) : I. The ubiquitous pollen: either on account of determination limitations or because the plant has a wide ecological amplitude II. The humid autochthonous: Myriophyllum cL spicatum, Typha, Polygonum persicaria Type, etc HI. The dry autochthonous: Amaranthaceae-Chenopodiaceae, Tamarix, Acacia, ere IV. Bryophytes and some local Pteridophytes V. Allochthonous of Mediterranean origin : Pinus, Quercus, etc VI. The southern allochthonous: a- humid: Caesalpinia Type b- dry: Randia c- afromontanous: Podocarpus, Hagenia cf. abyssinica. VII. Undifferentiated spores There are very few Pteridophyte species in Egypt. Most of the spores are allochthonous. However, the three generic determinations may indicate three ferns of Egyptian origin. Indicative of a local origin of Azolla is: variety of remains, fragile massulae tissue and their high concentration. Ophioglossum polyphyllum belongs to the Egyptian flora (Tackholm 1974). Pteris has several Mediterranean species. Zohary (1973) has observed the last two genera together in Cypriot meadows. Ophioglossum and Pteris could have been present in the delta. The significance of the Bryophytes is ambiguous. Riccia and several Anthocerotales species (three types here) thrive on temporarily humid areas such as river banks after floodings. They are classified here as autochthonous because they grow along delta branches even if some are allochthonous (their percentages increase when the percentage of allochthonous Pteridophytes increases). The classes of the synthetic diagrams are (Fig. 4): (1) Gramineae; (2) Cyperaceae; (3) Other ubiquitous and humid autochthonous; (4) Dry autochthonous; (5) Mediterranean and southern allochthonous. The Mediterranean allochthonous are always at low percentages, ca. 1%. Hence, there is no loss of information in including them in another class. Results Azolla determination The morphology of the reproductive organs of Azolla is complex (Perkins et al. 1985; Fowler et al. 1978). The sporophyte produces two types of sporocarps: the female megasporocarp and the male microsporocarp. The former contains a single megaspore (400 lxm). The megaspore is complex, composed of two parts : the spore itself (called the body) and the floats. In section Azolla Meyen there are three floats and nine in section Rhizosperma Meyen. The m a l e m i c r o s p o r o c a r p p r o d u c e s n u m e r o u s microsporangia (7 to 100). Each microsporangium is composed of three massulae or more; inside each of them there are 32 or 64 microspores (15-25 lain). The massulae are sometimes equipped with a harpoon-like structure, the 46 S6 $2 $7 S8 Depth in m Depth in m P.S. An An P.S. P.S. A n P.S. An 1-d- - Continental clay Centlnental c_U~%~l/IJ 183c zo • b-_ * Lagoonalor 2 ~805 _+ 4( i ....... Peaty clay 2340 + 9 i e4 1910 _+ 70 Lageonal clay 2-- 3 -- 1 . ........ d- - 3 Silt 4- - - P-e#t - - &# ~ I I ~I 380( -4 90 5-- Distdbutary mouth Sand bar ........ (~e~t!r~e dc - ~./.I l "i ! b- 6-- a- Delta front clayey-silt / at 5,50fn ;' // at 10.00 m Yr " • 3750 _* 60 Delta front Clayey-silt a.- 4695 4, 115 I zx /~ z~ z!, t Peat i-- ~olc~nic at 12.00 m herds st 14.80 rr~ ~; Legend 4230 _* 90 Prodelta mud a at 22.60 m - t ~ Roots c£3 Sand pocket ~-,~ Oxidized spots P.S. : Pollen sample A n :Aze//a ni/otica megaspore ( • high, * low conc~mlratlon) .... Bivalve shells Fig. 2. Sedimentologicallog and 14C ages of cores $2, $6, $7 and 88 glochidia or the trichomes, which serves to grip the female megaspores. The fossil megaspores and the microsporangia are found in the sand-size fraction and the microspores and massulae tissue in the palynological preparations. Based on the works of Martin (1976), Fowler et al. (1978) and Perkins et at. (1985), the best criterion to determine Azolla plants themselves is their reproductive apparatus, and more especially the megaspores and glochidia. Microspores have never been used to determine the species. They are very seldom described. The exine of A. nilotica is smooth on the Demalsi's drawings (1953) and in reference collection slides. A. africana Dew is illustrated in Maley (1970). This obsolete species mentioned on the herbarium sheet encompasses the two African species. On the basis of megaspore characteristics, it appears to be A. pinnata. The microspores are verrueate (Plate 1-2, 1-3). A. pinnata is illustrated by Straka and Friedrich (1989) where the microspores are shown as verrucate. We Fred the same difference in ornamentation in the reference slide collection. The $2.2 fossils are A. nilotica. A characteristic feature for this species is the glabrous exine of the megaspore body with distal spiny perinal excrescences (Plate 1-7). A. pinnata, on the other hand, has verrucae. The lack of simple glochidia or trichomes on the massulae is a secondary criterion for the species. The microspores are trilete, and are about 20 lxm (Plate 1- 4 and 1- 5). They have a thick wail with a completely smooth exine. These fossilized microspores are very similar to A. nilotica microspores. It seems possible to use the ornamentation as a species criterium. Sedimentological environment and stratigraphy Azolla megaspores have been recorded in two types of sedimentary environment (Table 1): either in continental clays between 1.00 and 4.10 m depth (samples $7.1, $8.5), or in marine sediments between 5.50 and 18.50 m depth (samples S13.27, S16.8). The concentration of fossils is very low in the marine sediments because of transport by Holocene distributaries and random scattering by the eastward sea currents. For example, sample S16.8 at 18.50 m, the megaspore has probably been carried by the Mendesian branch and deposited in the delta lobe while the branch was active. Sample in core Hm 30 000 W at 9.20 m was probably carried by the Tanitic branch. Samples S13.27 and $22.5 were brought by the Pelusiac branch before its decline 2000 years ago. In the continental clays the frequently higher concentration of megaspores and the presence of massulae soft tissue indicate very short transport. Freshwater marshes behind sand dunes have developed since the stabilization of the sea level and the consecutive fluviatile alluviation. Coastal lagoons and marshlands in Dynastic Times did not extend south of the present 2 m contour line. They were present as far back as 6000 B.P. (Butzer 1976, Fig. 4). Most of the cores are located below the 2 m contour line. 14C dates give ages varying between 5140 and 1830 B.P. All the samples with Azolla therefore belong to the second half of the Holocene, when arid climatic conditions similar to those of present day prevailed (Degens and Spitzy 1983). 47 1- 5 ' 'lO).trn 6 _ 7 j----.--~ 4~m 81 ! 5g,,um 9 i...._..,_,,_~ 33~m Plate 1. 1, Two microspores in massula tissue, sample $2c, Azolla cf. nilotica. Light microscope. 2, Microspores of Azolla africana, proximal view, trilete mark, reference slide 819 sp. - ORSTOM 68-85. Light microscope. 3, Microspores of Azolla africana, distal view, reference slide 819 sp. -ORSTOM 68-85. Light microscope. 4, Microspore in massula tissue, proximal view, trilete mark, sample $2c, Azolla cf. nilotica. Light microscope. 5, Microspore in massula tissue, equatorial view, sample $2c, Azolla cf. nilotica. Light microscope. 6, Spiny excrescences, distal part of the megaspore body, Azolla nilotica, US1755198. SEM. 7, Spiny excrescences, distal part of the megaspore body, sample $2-2, Azolla of. nilotica. SEM. 8, Megaspore, sample $2-2, AzoUa cf. nilotica. SEM. 9, Microsporangium or massula, sample $2-2, Azolla of. nilotica. SEM 48 Palynology Diagrams : description and interpretation (Figs. 3, 4) At least four palynological studies have been published on the emerged Nile Delta and Northern Egypt: Saad and Sami (1967) from the Berenberal region (Late Pleistocene and Holocene); Mehringer et al. (1979) from Birket Qarun (last 325 years); Sneh et al. (1986) from East of Suez Canal (Holocene) ; and Ritchie (1986) from Dakleh Oasis (modem spectra). Core $2. In samples S2a, S2b and $2c, there is a decrease in the percentages of the well represented Cyperaceae. Some peaks of one or another aquatic plant (S2a : Typha cf. domingensis; S2b • Myriophyllum cf. spicatum and T. cf. domingensis; $2c : Azolla cf. nilotica) are observed. The vegetation of the very dense papyrus marsh, which is perennial, seems to become more open. The spectra show no external influences such as from rivers (monolete psilate spores, tropical elements, etc) or from the sea (foraminifera, dinoflagellate cysts, etc). In sample S2d, the development of AmaranthaceaeChenopodiaceae and the presence of other dry autochthonous elements (Ephedra, Acacia, Tamarix) indicate the proximity of sand dunes or levee banks. The allochthonous grains with 12% (afromontanous elements from Ethiopian Uplands and spores of Pteridophytes and Bryophytes) mark the contact with the river, i. e. the Damietta Branch. The other micropalaeontological remains provide some information (Table 2). The high concentration of Rivulariaceae occurs in the peat level. It supports the idea of an isolated marsh in sample S2a, whereas in sample S2d, the Concentricystes are brought by the river. Roots between samples $2c and S2d indicate temporary emersion just after the highest concentration ofAzolla megaspares. In this core, it is possible to follow the progressive evolution of the vegetation from a very densely vegetated marsh completely isolated, ca. 3800 B.P., to a marsh with some open spaces, ca. 1830 B.P. Then, there is a temporary emersion which could correspond to the low flood levels during the Roman Period. Finally the site is under water again, probably as a river. It is edged by levee banks or sand dunes. The water might be slightly saline. Those successive developments might also have been influenced by the eastward migration of the Damietta Branch. The youngest 14C age is 1830 + 70 B.P. from sample $2-2. However 35 cm above, some megaspores are still present. The sediment is a continental clay and, sedimentation rate is ca. 1 mm per year. The last observation of Azolla therefore is around A.D.550 to 600, when Egypt is still part of the Roman Empire, the Arab conquest not yet having taken place. Special determinations Pollen and spores. The Cyperaceae is a large family that includes xerophytes and hydrophytes. In the Nile Delta, much of the Cyperaceae pollen corresponds in size and aperture number to that of Cyperus papyrus, especially in core $6 and samples S2b, $2c and S2d. Wild papyrus progressively disappeared from Egypt during the last centuries (Tltckholm and Drar 1950). Other Cyperaceae are also present for example in core $8 and in samples S7b and $7c. Amongst the Gramineae, it is impossible to distinguish different genera. Cereales might be present considering the antiquity of agriculture in the delta ca. 5000 B.C. (Butzer 1976). It is impossible to separate the pollen of cultivated versus non-cultivated grasses in North Africa (Maley 1981). The bulk of the pollen grains might be Phragmites. It is often found growing in association with papyrus where it forms a Phragmitetea community. The Typha monads might be T. domingensis Pers. It is the only Typha which produces monads in Egypt today. Its present day habitat is marshes and running waters. Some tetrads have also been found. They may represent T. elephantina Roxb. which occurs in marshes only. This species seldom occurs in present day Egypt. Other micropalaeontological remains and their ecological significance (Table 2). The type numbers given below correspond to Van Geel's types. Some microfossils only indicate freshwater: Ceratophyllum leaf-spines (Ceratophyllaceae), cf. A m p h i t r e m a (Type 3, Thecamoebae), Pediastrum (Euchlorophyceae). The presence of Zygnemataceae spores is indicative of stagnant, shallow and more or less mesotrophic freshwater habitats: Mougeotia (Type 61, Type 313), Debarya, Spirogyra (Type 130, Type 315). The Rivulariaceae (Type 146, sheaths of Cyanophyceae) mark low concentrations of nitrogen and phosphorus, eutrophy, alkaline environment and the presence of a lot of oxidizable organic compound. The incertae sedis, Concentricystes, is considered as indicative of water runoff and rivers. The presence ofAzoUa massulae soft tissue and of pollinia or fragments of them is only possible where there is minimal transport of material. Other types recorded include: Tetraploa (conidia of a fungus, Type 89), a vasiform microfossil (Type 179), O a d ~ r a (Type 72, Arthropoda). Core $6. All the spectra are largely dominated by Cyperaceae. We note a very slight regression of papyrus in favor of Gramineae, probably Phragmites. The presence of Polygonum persicaria Type and Typha cf. elephantina is continuous whereas there are no palynomorphs of allochthonous origin, nor Bryophytes, or Pteridophytes except Azolla microspores and megaspores in sample S6d. The four samples give no indices of external influence. This means that they may not be receiving water from floods. The concentration in palynomorphs is very high: about 49 Ephedra Acacia sect. Acacia × Combretaceae I Tamafix Peristrophe cf. bicalyculata .< Amaranth.-C he nopodi~aceae ~ "I ""I cn m Aerva Calystegia Cucurbitaceae Thymeleaceae X Anthocerotales T.1 Anthocerotales T2 n m n I Anthocerotales m~ . . . . " I naB IWx ~ x ~ " --~TT "-~I ~: or;!oT::L:,. I,'1I!1 Ili, + 5 ~m o ttl,tRi-,l ,,,, Labiatae Ocimum Quercus Mavaceae peril:era e Abutilon m Caesalplnia Type Syzygium Typ~ Sapotaceae Mimusops MonocotyL monocolpate Papilionacoae PIantaginaceae J Umbelliferae _~ m Loranthaceae 1 Randla Podocarpus ~ x x x Hyphaene Type Cardiospermum halicacabum Myriophyllum spicatum ~ Nymphaea Onagraceae Hagenia abyssirtica Monoletes (smooth) mime- i m I l m 7 i- x x Ir" " . Triletes( . . . . •.I~--~ IT--T i Tdletes (not smooth) i -iT • """ r'll i I b Ericaceae Cannabaceae ' rrl Ligullflorae roo "1 Olea Tilia Ulmus "lml 1:2 Cruciferae Oleaceae X rn Attemisia Centaufea Gramineae Picea Pinus Betula Alnu~ Sambucus nm ~ x T3 Pinaceae x ~ r-" TubuIiflorae _~ o Azolla nilotica ~-In x x ~---T Riccia i T ~ I~ • Cyperaceae Caryophyllaceae m--T7 zw Boraginaceae -I ~" i Rumex x Monoletes {not smooth) Potamogeton Ranunculaceae th) ~ n Undetermined "r'I Undeterminable a- Polygonurn pers[cada T. J ~I "' II'- Typha cf. domingonsis Typha cf. elephandna Fig. 3. Detailed palynological diagrams 80 000 grains/g dry sediment. This represents the very high local production. The large number of Rivulariaceae and Zygnemataceae indicates stagnant water. The sediment is peaty except sample S6d which is clayey. There is a 14C date of 3750 + 60 B.P. between samples S6a and S6b (under the Azolla-rich level). Above theAzolla-rieh level, at 2 m depth, there is a date of 1910 + 70 B.P. In conclusion, the samples give the image of a pure dense papyrus marsh with locally open water flee for the development of Azolla. Papyrus usually occupies the centre of the marsh area and other plants grow only on narrow margins around it. As the percentage of the other plants increases, it is inferred that the Size of the marsh decreases. The marsh seems completely isolated from rivers and the sea. It is far inland between two major distributaries" the Mendesian and the Tanitic branches. The Tanitic branch was initiated at about 3150 B.P.. The Mendesian branch already existed at 8000 B.P.. It was partially silted up by 2450 B.P. The archaeological site of Tanis was occupied around 3100 B.P. We know from archaeological evidence that there were extensive areas of lagoons and marshes 50 $2 P.S. Depth in m $7 An LEGEND ~ ] Gramineae~ F~ Cyperaceae 1 d ~ ~ utochlhonous humid Autochthonous dry c 2 AIIochthonous (2 classes) $8 25% 50% 75% 100% $6 5.0 \\~ \ \ . ... ... . .. . .. . .. . ... ... ... . .. . .. . . . . . . . . . . . . .- . .~ . . - . . - . .- . .- . .~ . . ~. . ~. . -. . . 6.0 k:~-----_:-:~i~i i i i i i~i~i~i~;ii r---- ---- - - - - - - - : 6.5 . . . . • . . . . . V .... I P, S. : Pollen sample A n :Azol~ ngot~a megas~re 25% 50% 75% iiii:i:i:i:i:i:i:i:i:i: 100% 25% 50% 75% 100% ( • high, ~ b w ~ n c e n t r a l i o n ) Fig. 4. Synthetic palynolog~ diagrams around the settlement. Land reclamation, swamp drainage, irrigation systems were already known long before the foundation of the Egyptian city of Tanis (Butzer 1976). Core $7. In sample S7a, in which there is a high diversity of taxa, all the classes are well represented. There are many Pteridophytes spores, R i c c i a spores and some afromontanous elements. In samples S7b and $7c, we register a strong influence from papyrus and cane marsh, and Azolla is well represented. The percentagt, s of AmaranthaceaeChenopodiaceae point to the proximity of sand dunes or river banks. The low number of allochthonous palynomorphs indicates a slight influence of flooding. Soon afterwards, in samples S7d and S7e, papyrus no longer occurs. Gramineae and AmaranthaceaeChenopodiaceae reach a high percentage. Some Onagraceae are present. There are very few southern allochthonous palynomorphs. The concentration in palynomorphs is much lower: about 10 000 grains/g dry sediment. The environment is quite aerobic. Concentricystes in sample S7a are brought by the fiver and the Foraminifera in samples $7c and S7b by the sea. Sample S7a is 40 cm over a volcanic ash layer. This may be Upper Minoan ash from Santorini, dated ca. 3500 B.P. (Stanley and Sheng 1986). This is not in agreement with the 14C date at 2.70 m of 3805 + 40 B.P. (Foucault and Stanley 1989) or with the palaeogeographic maps of Coutellier and Stanley (1987). A 14C date at 1.50 m gives 2340 + 90 B.P. Sample S7a comes from the delta front clayey-silt, the four other samples from continental clay deposit in marshes or lagoons. The roots between samples S7d and S7e originate from a temporary emersion. The evolution of the delta from sea to continental lagoon or marshes is illustrated in $7. Areas favourable to Azolla are near and in temporary contact with the sedimem studied. Core $8. Samples S8a and S8b come from the prodelta mud. All the classes are represented and especially the allochthonous Ptefidophytes, up to 18%. There are some afromontanous elements, some Foraminifera and Concentricystes. Samples $8c and S8d have similar spectra. Sample S8d, however, has more local freshwater elements and less allochthonous. Azolla megaspores are present here and there. The sediment is silty. The material belongs to fiver beds. Azolla f'mds are younger than 4230 + 90 B.P. Here it best appears that few pollen grains come from the desert vegetation. The percentages of AmaranthaceaeChenopodiaceae and other dry elements are low while, at the same time, there are relatively high percentages of southern allochthonous elements and river-transported spores. Discussion and causes of palaeoenvironmental changes leading to demise of Azolla The percentage of autochthonous versus allochthonous palynomorphs is always very high (ca. 90%) except in samples from the prodelta mud. Spectra with little or no allochthonous palynomorphs indicate isolation from the 51 Table 2. Microfossils and some pollen data m ~3 o $2 ,~ • rJ d c b a $6 c .cz + + + + + E'~ b-- < * * + (+) + g Fungi 360 271 308 673 Fungi and 550 386 487 862 * d + * + c + * * + b + + a + * * * * (+) + CP_p~cutophyllum $7 e + d + c S8 + + + + b + + + a + + + d + + + + + C + + + + + + b a + + + + + + + + + + + * + + + 438 392 262 249 597 (+) (+) + + + + DeboJ~ya + N Cladocera and Ceratophyllum 540 433 285 230 ~ + +. + 70,512 83,289 91,722 7,778 4,857 14,171 + + + + + + + + + + + + * high concentration + medium concentration (+) low concentration (1) basic sum of pollen and spores used in detailed diagram (2) concentration of pollen and spores per g of dry sediment Nile River and from the sea. The growth of the Nile Delta and the accumulation of sand dunes isolated numerous perennial freshwater marshes with pure papyrus cover. On the other hand, high percentages of allochthonous palynomorphs mean contact with floods, fiver, sea, etc. Most of the allochthonous material comes from Tropical and Equatorial Africa. Some taxa can only originate from the Ethiopian Highlands. Azolla-rich levels correspond to well developed papyrus marshes. In Israel, 4000 years ago, papyrus marshes in the Hula basin were widespread (Bein and Horowitz 1986). Even nowadays, the plant list, according to Zohary (1973), includes Cyperus papyrus, Phragmites australis, Polygonum acuminatum, Nuphar luteum, Typha, Scirpus, Inula, Sparganium. At the time of the Pharaohs, Egypt might have been at the boundary of the natural distribution of Azolla. The vegetation of the Nile Delta might have presented some similarities with that of the Hula basin. Some plants have become rare or have disappeared from Egypt: papyrus, Azolla, Typha elephantina, Myriophyllum spicatum, etc as a result of reduction of freshwater marshes behind coastal sand dunes. The reasons for this may include the following: 1. The climate seems to have been hyper arid for the last 4000 years (Degens and Spitzy 1983). There has not been a major climate change during the last two thousands years, which would coincide with the disappearance of Azolla, after A.D. 600. Climatic change cannot, therefore, be a direct cause of the disappearance of Azolla, only possibly a long-term one. 2. Prolonged low floods in the delta are a consequence of the climatic conditions upstream in the Tropics and especially the decrease in summer monsoon rain on the Ethiopian highlands. Flooding is due to the Blue Nile. The White Nile maintains a much more constant flow corresponding to a less marked Equatorial seasonal variation. Its input is measured by the low fiver levels in Egypt. Geomorphological studies, and archaeological and historical records for the last 5000 years have produced a very complete curve of the Nile level changes in the delta. Some well documented low level periods are 2100 B.C., 1200 B.C. and the Roman Period. The two oldest correspond to the two Dark Ages in the period of the Pharaohs (Bell 1971). However, low flood levels alone cannot be the reason for the disappearance of some species belonging to the marsh vegetation. 3. The salinity has increased due to the delta subsidence and there has been a slow rise in sea level (Stanley 1988). Those two factors could only explain the displacement of the marshes. 4. A much more likely factor is human activity on a broader scale. Agriculture here is ancient and the concentration of population has always been high. Azolla 52 has been regarded as a weed because it obstructs the irrigation channels. Its habitat has been reclaimed for more arable land. Conclusion The megaspores and microspores of A. nilotica have been found outside its present area of distribution in recent Holocene sediments of the Nile Delta. A new criterion for species determination has been found in the microspore ornamentation to differentiate A. nilotica from A. pinnata. The former is smooth while the latter is verrueate. Through palynological studies we could establish that the environment appropriate to the fern has disappeared from Egypt. It was present up to the Roman Period. Its disappearance is most likely attributable to human activity. The modern Egyptian rice paddies could benefit from the re-introduction of Azolla, a green manure. This fern was once part of the vegetation of the Delta. Acknowledgements.The project has been assisted by two short-time visit grants of the Smithsonian Institution (S.I.), Washington, D.C., in 1986 and 1987. I wish to express my gratitude to Dr. F. Hueber of the S. I. for inspiring the paper. Thanks are also due to Dr. D. J. Stanley, S. I., for providing invaluable samples. The research was carried out with the collaboration of Bill Boykins, Laboratory of Sextimentology, S. I.; Prof. G. Seret of the laboratoire de Pal~.og~ographie et Pal6ontologie, UCL, Belgium; and finally the Laboratoire de Palynologie, Montpellier, France. References Bein A, Horowitz A (1986) Papyrus - a historic newcomer to the Hula Valley, Israel ? Rev Palaeobot Palynol 47:89-95 Bell B (1971) The Dark Ages in Ancient History. I. The first Dark Age in Egypt. Am J Archaeol 75:1-26 Butzer KW (1976) Early hydraulic civilization in Egypt, a study in cultural ecology. The University of Chicago Press, Chicago Coutellier V, Stanley DJ (1987) Late Quaternary stratigraphy and paleogeography of the Eastern Nile Delta, Egypt. Mar Geol 77:257-275 Degens ET, Spitzy A (1983) Paleohydrology of the Nile. Mitt Geol- Paleont Inst Univ Hamburg (SCOPE / UNEP Sonderband) 55:1-20 Demalsy P (1953) Le sporophyte d'Azolla nilotica. Cellule 56:1-60 Dricot E, Leroy SAG (1989) Peptization and sieving for palynological purposes. Geobound, Brussels, 2:114-126 Foucault A, Stanley DJ (1989) Late Quaternary paleoclimatic oscillations in East Africa recorded by heavy minerals in the Nile Delta. Nature 339:44-46 Fowler K, Stermett-Willson J (1978) Sporoderm architecture in modern Azolla. Fern Gaz 11:405-412 MacLeay KNG (1955a) Geographical relationships of the Pteridophyte flora of the Sudan. Bull. Jardin Bot l'Etat Brux 25:213-220 MacLeay KNG (1955b) A preliminary list of Pteridophytes of the Anglo-Egyptian Sudan. Webbia 11:587-606 M~ey J (1970) Contributions h l'6tude du Bassin tchadien. Atlas de pollens du Tchad. Bull Jard Bot Nat Belg 40:29-48 Maley J (1981) Etudes palynologiques dam le bassin du Tchad et pal~oelimatologie, de l'Afrique nord tropicale de 30000 am ~ l'~poque actuelle. Th~se So., Univ. Montpellier. Tray. doe. de rORSTOM 129. Paris Martin ARH (1976) Some structures in Azolla megaspores, and an anomalous form. Rev Palaeobot Palynol 21:141169 Mehringer PI, Kenneth L, Petersen L, Hassan FA (1979) A pollen record from Birket Qarun and the recent history of the Fayum, Egypt. Quat Res 11:238-256 Moore AW (1969) Azolla: biology and agronomic significance. Bot Rev 35:17-34 Perkins SK, Peters GA, Lumpkin TA, Calvert HE (1985) Scanning electron microscopy of perine architecture as a taxonomic tool in the genus Azolla Lamarck. Scanning Electron Micrnse 4:1719-1734 Ritchie JC (1986) Modern pollen spectra from Daldaleh Oasis, Western Egyptian Desert. Grana 25:177-182 Saad SI, Sami S (1967) Studies of pollen and spores content of Nile Delta deposits (Berenbal Region). Pollen Spores 3:467-503 Sneh A, Weissbrod T, Ehrlich A, Horowitz A, Moshkovitz S, Rosenfeld A (1986) Holocene evolution of the northeastern comer of the Nile Delta. Quat Res 26:194206 Stanley DJ (1988) Subsidence in the Northeastern Nile Delta : rapid rates, possible causes, and consequences. Science 240:497-500 Stanley DJ, Sheng FI (1986) Volcanic shards from Santorini (Upper Minoan ash) in the Nile Delta, Egypt. Nature 320:733-735 Stanley DJ, Sheng H, Pan Y (1988) Heavy minerals and provenance of Late Quaternary sands, eastern Nile Delta. J Aft Earth Sc 7:735-741 Straka H, Friedrich B (1989) Palynologia Madagassica et Mascareniea. Familien 1 bis 16, PteridoptlytaGeneralindex. Akad. Wissensch. und Liter.-Mainz-72. Steiner, Wiesbaden T~ckholm V (1974) Students' flora of Egypt. Cairo University, Beirut Tiickholm V, Drar M (1950) Flora of Egypt. Fouad I Univ. Press, Cairo Van Geel B (1978) A paleoecological study of Holocene peat bog sections in Germany and The Netherlands. Rev Palaeobot Palynol 25:1-128 Zohary M (1973) Geobotanical foundations of the Middle East. Fisher, Stuttgart