Pollen analytical and geological investigations of the Lower and Middle Pleistocene in the Northern Netherlands. The Pleistocene deposits in the northern part of the Netherlands, form through their extreme thickness up to more than 300 m, a promising object for study from the stratigraphical point of view. The boulder clay of the penultimate glaciation lies in the province of Drente at or near the surface, but in northerly and westerly directions it dips away beneath Upper Pleistocene and Holocene deposits. The Pleistocene deposits below this boulder clay form the subject of the present paper. In the first chapter the existing opinions regarding the stratigraphy are briefly reviewed. Van Cappelle (1888, 1891, 1892a, 1892b) and Lorié (1887a, 1893, 1899) described several borings already in the last century. They found the boulder clay in general to be underlain by a fine grained sand below which a coarse sand occurred, containing many Scandinavian erratics. Both, Lorié and Van Cappelle considered the coarse sands as having been deposited under the simultaneous influence of southern rivers and northern fluvio-glacial streams. Therefore it was assumed that a Scandinavian ice-front was at that time not too far distant. Van Cappelle believed that this ice-front belonged to an older glaciation than that from which the boulder clay originated, since he found in the fine sediments some rare plant remains which indicated a temperate climate. Lorié, however, took the ice-front to be a mere oscillation of the same glaciation which deposited the boulder clay. Later geologists agreed with Lorié (Tesch, 1934, 1937, 1947; Steenhuis, 1939), basing their opinion on a correlation with the central part of the Netherlands, where a coarse zone occurs above the Nede clay representing the Mindel-Riss Interglacial (Florschütz and Jonker, 1942). It is at present generally accepted that these coarse gravel containing sands form one continuous horizon which was deposited during the first cold stage of the Riss glaciation, whereas the boulder clay represents the second cold stage. Hardly anything is known about deeper horizons in the northern Netherlands. According to Tesch and Steenhuis Mindel-Riss Interglacial deposits are lacking; the Mindel Glacial stage is perhaps represented by an other zone of coarse deposits overlying again a series of fine grained sediments. The base of the Pleistocene is formed by marine beds which, in the opinion of Tesch (1934, 1937), comprise in part, the Günz Glacial stage, since the molluscan fauna of the middle part bears a distinctly arctic character. Since the stratigraphy was based on lithological criteria alone it seemed desirable to investigate whether it could be confirmed on palaeontological grounds. For this purpose pollen analysis proved to be the most suitable method, since the sequence of sediments under review is accessible only in borings, which seldom yield macroscopic fossils. In chapter II some technical aspects of the pollen analytical research are discussed. The material available consisted of samples of some rare peat beds, of samples of often thick clay beds and of lumps of clay and peat, which are found sometimes in sand samples. Such lumps probably originate from very thin peat beds, which were not differentiated in the sample as separate layers. This is proved by the similarity between the spectra of these lumps and the spectra of thin beds found in situ in nearby borings (table II and III, p. 272). In general very sandy samples were not analysed. unless they were very humic and nearly sandy peat in appearance (table IV, p. 273). The samples investigated proved to be rich enough in pollen to furnish reliable counts. A number of absolute pollen frequencies of different materials are given in table I (p. 271). All the samples were prepared after the technique introduced by Erdtman (1943). Some spectra indicated pollen of tertiary genera, which became extinct in western Europe at the beginning of the Pleistocene period. These pollens are clearly derived from tertiary deposits, but the same may be the case with other pollen of less diagnostic character. The question thus arises to which extent the pollen content of the strata under study has been derived from older series. The correction technique developed by Iversen (1936), who subtracted the clearly secondary pollen he found in a boulder clay from the spectrum he obtained in the complex overlying it, could not be applied in our case, where the sediments studied unfortunately underly the boulder clay. It might be feasible to correct contaminated spectra by comparing them with pure spectra from nearly the same depth, so that possible climatic differences can be disregarded. Contaminated spectra were found in several thick clay complexes. In these cases intercalations of peat beds which could have yielded uncontaminated material for comparison do not occur. Therefore the amount of contamination could not be estimated and all spectra given are uncorrected. However, we have indicated in the diagrams the amount of typical tertiary pollen (T) and of Hystrix (H), expressed in percentages of the tree pollen. However, four arguments indicate that the amount of reworked pollen in general not great: 1. the low frequency of Hystrix, which proved to be a measure of the impurity in Iversen’s material; 2. the absolute pollen frequencies of the impure spectra as compared with the pure ones; 3. the very small pollen content of the boulder clay; 4. the similarity of pure and impure spectra in corresponding zones of different diagrams. The depth given for each spectrum in the diagrams corresponds with the average depth of top and bottom of the sample (in metres below N. A. P. = high water at Amsterdam). In chapter III the results of the pollen analysis are discussed, whereas in chapter IV a stratigraphical interpretation is attempted after comparing the diagrams. The number of Lower and Middle Pleistocene pollen diagrams from western Europe is still very limited. The most important diagrams are those from Quakenbrück (Wildvang, 1935; Jonas, 1937a) and Ummendorf (Selle, 1941), both from western Germany. The diagrams from Starup and Harreskov, published in the classical paper by Jessen and Milthers (1928), have not been used since it is quite uncertain whether the penultimate interglacial referred to by the authors can be correlated with the Mindel-Riss or the Riss-Warthe Interglacial. Of the borings investigated (for localities compare map, fig 1) Bantega yielded by far the best diagram. This boring has specially been carried out for this purpose and yielded a complete sequence of undisturbed cores. The diagram is found to be in close agreement with those of Quakenbrück and Ummendorf: during the hardwood phase the mixed oak forest reached its maximum at an early date (20.65—20.85m); only afterwards Carpinus and Abies appeared and Picea had a distinct (double) maximum after the climatic optimum of the hardwood phase (18.54 and 17.94 m). There can be no doubt as to the Mindel-Riss Interglacial age of this diagram. The same interglacial epoch can, based on a smaller number of spectra still clearly be recognized in some other diagrams viz. Bergumerheide (41.60—65.0 m), Sneek (24.0—46.0 m), Spannenburg (20.40—88.0 m), Lemsterland (15.58— 54.47 m) and Gasselte (27.80—62.65 m). Furthermore Bergumerheide (7.00 m) and Spannenburg (9.30 m) show a temperate spectrum close below the boulder clay, representing probably the Riss I/II Interstadial stage. Three deeper borings again show beneath the Mindel-Riss Interglacial a number of spectra with a temperate character. In a few of these pollen of Pterocarya occur (Spannenburg, 211.90 to and deeper; Lemsterland 153,40 m). Pterocarya has long been known from the Tegelen clay, but the age of this famous locality has not yet been determined with certainly. Tesch (1934, 1937) and Florschütz (1939) hold the view that it represents the Günz-Mindel Interglacial, but at present some authors consider the Tegelen clay as belonging to the Günz I/II Interstadial. Our palaeontological knowledge of the Lower Pleistocene is still too incomplete to solve the problem. We do not know whether some so-called tertiary relics (for inst. Tsuga, Pterocarya) which occur in the Tegelen clay, range upward into the first interglacial or not. In this respect we may remark that the diagrams from Spannenburg and Lemsterland possess no indications of cold spectra which could he interpreted as Günz II Glacial stage. An equivalent of the Tegelen clay has probably been found at Bergumerheide. Florschütz (1938) mentioned Azolla tegeliensis (155—157 m) which he suggested as a characteristic species of the zone of Tegelen. At Spannenburg another species (A. filiculoides) which is characteristic for the Nede clay (Mindel-Riss Interglacial) has been found between 25 and 40 m. A boring near Dordrecht however, yielded both species from the same bed. Bergumerheide shows between the horizon with A. tegeliensis and the Mindel-Riss Interglacial deposits a third zone with spectra of a temperate climate and without tertiary relics (88.0—125.25 m). This strongly suggests that the zone of Tegelen belongs to the Günz I/II Interstadial and that the abovementioned horizon at Bergumerheide represents the Günz-Mindel Interglacial. Perhaps the spectra between 187.15 and 197.90 m of Spannenburg belong to the same horizon. An entirely different picture is shown by the diagrams of Assen and Winschoten. Pinus predominates throughout the diagrams, but nearly all spectra are contaminated with tertiary pollen and with Hiystrix. As we do not know to which degree the percentages of certain genera are overrated, the diagrams are less valuable. The diagram of Assen is particularly monotonous. Winschoten shows rather high Alnus percentages in the lower part. The same has been observed in a thin peat bed of the boring Zuidbroek (93.17—93.27 m, table V, p. 291), which possesses 2% Tsuga pollen. Since in this case we have no reason to believe that the sample is contaminated, it must be assumed that the entire pollen content is autochthonous and therefore the spectrum probably represents the (Günz I/II Interstadial. The same horizon can perhaps be recognized in the nearby boring at Winschoten. The same picture as at Assen and Winschoten is shown by the diagram of Drouwen and the lower part of Sneek. We shall see later that geological considerations are of assistance in interpreting these profiles. In chapter V further consideration is given to the Middle Pleistocene marine horizon and a new conclusion has been reached regarding the age of this deposit which has a hearing upon the general understanding of the glacial stratigraphy of the northern Netherlands. In two borings the Middle Pleistocene sediments are partly developed in marine facies (Bergumerheide 46.00—62.00 m; Sneek 31.00—42.00 m). The marine deposits lie above a coarse fluviatile series which is generally held to be of early Riss Glacial age. The pollen diagrams prove the marine sediments to be deposited between the mixed oak forest phase and the Picea phase, i. e. in the second half, of the Mindel-Riss Interglacial. The underlying coarse deposits must therefore have been laid down during the first half of the same interglacial. A Mindel-Riss Interglacial transgression is known from England as well as from Germany. In East Anglia the Corton Sands (Baden-Powell and Reid Moir, 1942; Zeuner, 1945) and in N. W. Germany the sediments of the Holstein-See (Grahle, 1938) were deposited during this transgression. The molluscan fauna (Tesch, 1939) proves to be somewhat colder in character than the landflora which is understandable since the sea transgressed from the north and consequently introduced northern species. On the other hand free immigration from the south was possible by land. Reid (1890) has already pointed out that different conclusions may be reached regarding climatic conditions from a comparison of marine and terrestrial organisms. Furthermore several rivers are known to have had their main aggradation phases during interglacial times when the rising sea level decreased their transporting capacity and thus were forced to deposit their load. As Zeuner (1945) stated we have to distinguish between thallassostatic terraces in the lower and climatic terraces in the middle courses. All classic studies of river terraces have been made of climatic terraces with glacial aggradation and interglacial valley formation. The lower courses of two west European rivers have been studied from the eustatic point of view: the Somme (Commont, 1910; De Lamothe, 1918; Breuil and Koslowski, 1931/32) and the Thames (King and Oakley, 1936) and the results are briefly reviewed. It seems that the same conditions which obtained in the lower Somme and the lower Thames were present in the northern part of the Netherlands: the principal factors affecting the behaviour of the rivers were oscillations of the sealevel during glacial and interglacial times, although terraces in the morphological sense did not develop owing to the gradual subsidence of the North Sea basin. Since all relevant deposits are entirely covered by younger sediments we do not yet know where in this country the transition occurs from the type of sedimentation typical for the lower course of the rivers under predominant marine influence to the climatic terraces of their middle course. Three marine transgressions are known to have occurred in the Netherlands Pleistocene: 1° the Eemian, of Riss-Würm Interglacial age which has long been generally recognized; 2° the Middle Pleistocene transgression, the Mindel-Riss Interglacial age of which has now been proved by the pollen diagrams; and 3° the Lower Pleistocene transgression (Icenian), the exact age of which is still unknown. Therefore the question arises whether this oldest transgression might be connected with the first interglacial. This point is discussed in chapter VI. Tesch (1934, 1937) argued that the middle part of the marine Lower Pleistocene represents the Günz Glacial stage on the ground that its molluscan fauna is distinctly arctic in character. Unfortunately the oldest marine deposits occur in only one of the borings investigated (Lemsterland). A few spectra immediately above this horizon show a temperate character. Though the horizon from which the spectra are derived may possibly be separated from the marine deposits by a stratigraphical hiatus, we must consider the possibility that the marine Lower Pleistocene horizon has not been deposited under such cold conditions as would appear from the conclusions of Tesch. The lowermost spectrum shows a small amount of Pterocarya pollen, suggesting, in accordance with our present state of knowledge, a Günz I/II Interstadial age rather than a Günz-Mindel interglacial age. However, if Pterocarya would still prove to have occurred in Günz-Mindel Interglacial time, the marine Lower Pleistocene could be ascribed to a Günz-Mindel Interglacial age. Incidentally, during this epoch, a high sea-level has been observed all over the world. The map (fig. 2) shows the location of all borings of more than 150 m deep, whereas some other important borings have been added in the southern part of the area. The following is an explanation of the figures shown on the map with each boring: if two figures are given the first means the depth to the top of the marine Lower Pleistocene in metres, whilst the second, in brackets, indicates the lowest level reached without penetrating this marine horizon. One figure in brackets indicates the greatest depth reached. In this case no marine Lower Pleistocene or older formations have been met with. A figure preceded by T means that the Tertiary has been reached at this depth without encountering marine Lower Pleistocene. The map shows that the Lower Pleistocene in a marine facies is restricted to the western part of the area. Since the base of the marine beds has nowhere been reached we do not know whether it is transgressive, as it is in the southern Netherlands. Comparing the figures of Zwartsluis, Vollenhove and Lemsterland with that of Spannenburg, there is a striking difference which suggests that the marine horizon has disappeared by erosion in the sub-soil of Spannenburg. The profile of Spannenburg gives no evidence of the relative age of the sediments present as compared with the marine deposits elswhere. It follows from the above that the Pleistocene of the area under investigation is not composed of a number of continuous horizons laid down one upon the other. Up to now it was assumed that coarse sediments represented cold or glacial periods, whereas finer deposits corresponded with wanner and intraglacial phases. However comparing the results of the Spannenburg and Lemsterland borings in this light, we are struck by the fact that the deposits of the Mindel Glacial phase at Lemsterland are composed of coarse grained sediments, whilst the equivalent interval at Spannenburg is fine grained. Both show, however, spectra indicating a cold climate. Furthermore at Bergumerheide a continuous coarse section is formed which shows a diagram containing successively warm, cold and again warm spectra, which correspond with the Günz-Mindel Interglacial, Mindel Glacial and Mindel-Riss Interglacial respectively. This demonstrates that grain size alone is not indicative of the climatic conditions prevailing during sedimentation. A number of borings have been investigated from a sedimentary penological viewpoint by Edelman (1933) and by Böhmers (1937). From a stratigraphical viewpoint, the B-Scheemda group is most interesting. It is characterized by high percentages of para-metamorphic minerals, the amount of which decreases from east to west. Edelman therefore looked for their origin in an easterly direction. Later he suggested that this group might have been deposited during the Mindel glaciation, when the courses of German rivers were deflected through the northern Netherlands by the icefront (Edelman, 1939). In the boring at Urk the most conspicuous B-Scheemda influence occurs between 66 and 99 m; in the Kippenburg boring between 66 and 92 m, whilst the nearest wells of Lemsterland and Spannenburg show spectra with a cold character at corresponding depths. In the Suameer boring the B-Scheemda influence is, according to Böhmers, most distinctly developed between 163 and 175 m. However, his mineralogical table (Böhmers, 1937, p. 62) shows the presence of two more horizons with rather high percentages of para-metamorphic minerals, viz. between 122 and 142 m and between 70 and 80 m. The lowest interval could not be investigated for its pollen content. The 122—142 m interval is unfortunately barren of pollen but the underlying and overlying beds contain a temperate flora. The uppermost zone (70—80 m) is apparently equivalent to the Mindel Glacial interval as established at nearby Bergumerheide. Thus it is obvious that the B-Scheemda group or a mineral aggregate comparatively rich in parametamorphic minerals fluctuated at different times and that these fluctuations might correspond to periodic deflections of the Germanic rivers into the northern part of the Netherlands. A discussion of the considerable clay deposits, which puzzled Lorié already half a century ago, is given in chapter VII. These deposits are known from a limited number of localities. In general all fine grained sediments between the boulder clay and the upper coarse horizon have been considered to form one unit, comprising the Riss I/II Interstadial as well as the early fluvioglacial deposits. The pollen diagrams strongly indicate that this cannot be the case. The clay deposits of Bantega and the upper clay of Spannenburg are undoubtedly of Mindel-Riss Interglacial age. Assen and Winschoten, however, represent another, probably glacial type, also by their content of reworked pollen. Of further localities known to have very thick clay beds Dronrijp is the most interesting. Here the clay is clearly deposited in a valley cut in the marine Mindel-Riss Interglacial series and therefore is apparently younger. As the boulder clay shows hardly any depression above the clay, the valley must have been filled up before the Riss ice reached the region (pl. XLVI). All localities with clay deposits thicker than 70 m, including occasionally fine sand beds, have been shown in black on the map (fig. 3). The two figures given indicate the top and bottom of the clay horizon, unless the second figure is in brackets, which means that at that depth the base has not been reached. Their distribution suggests that the localities belong to two valleys, one running from east to west and a second one from southeast to northwest. The thick clay deposits in the lower half of the Sneek boring correspond, from a pollen analytical point of view, to the clay deposits of Assen and Winschoten, which are of a post Mindel-Riss Interglacial age. The clay of Sneek however, is covered by the marine Mindel-Riss Interglacial beds and must therefore be older. This suggests that the thick clay beds of other localities might also have been deposited in two phases, although separating interglacial beds are lacking in most of the profiles. The clay itself is an unpromising medium for pollen analysis, but better results may be obtained from the border regions where intercalations of sands with peat might occur. The profiles (fig. 4) show a gradual thinning out of the coarse deposits in the direction of Assen. The profile at Drouwen, in the vicinity of Gasselte, yielded spectra with a cold character and with reworked pollen above (see plate XLV) as well as below (refer table, p. 290) the coarse horizon. Strong evidence therefore exists that thick clay beds were deposited twice in valleys of glacial age during the Lower and Middle Pleistocene. The older of the two deposits apparently correspond with the „Lauenburger Ton” from northwestern Germany (Schucht, 1912). Apart from the above discussed valley systems of Sneek and Dronrijp a still younger valley system is known to exist. It developed after the above mentioned two valley systems had been filled up and it still existed at the time the Riss ice arrived because we find a mantle of boulder clay deposited in the bottom as well as on the flanks of these valleys. Such valleys have been recognized below the present Overijselse Vecht and the Ems rivers. Essentially we have recognized three periods during which valleys were formed in Middle Pleistocene times. The first (Sneek) is pre Mindel-Riss Interglacial and therefore probably of Mindel Glacial age. The second (Dronrijp) and the third (Vecht-Ems) both occur between the Mindel-Riss Interglacial and the arrival of the Riss ice and therefore probably correspond with the two cold stages of the Riss Glacial. From this scheme it might be concluded that the Riss ice reached the Netherlands during the Riss II stage. Chapter VIII summarizes in Dutch the sequence of events during the Lower and Middle Pleistocene and we refer to table (p. 333) for a chronological representation of most of the above described events.