In the introduction and chapter 2 the incentives and way of reasoning are given for the description of an evolutionary basis of pollination ecology. Starting from the until recently rather anecdotical character of the study of pollination ecology as a whole, and in the absence of large-scale correlations of flowerecologically important character states with angiosperm and insect phylogeny (in the sense of Hennig, 1966), an attempt is made to derive directed evolutionary lines (transformation series) of floral character states. Because the fossil record of flowers is very restricted and the study of angiosperm phylogeny in the sense of Hennig (1966) is only at its very beginning, the only possibilities to achieve this, can be based on the comparatively well-known fossil history of insects and insect phylogeny, the least anecdotically compiled survey of insect visits to flowers, whether resulting in pollination or not, by Knuth (1898a and 1899) as far as it concerns the central European area (most thoroughly known area concerning insect fauna and angiosperm flora), and the fossil record of extant angiosperm taxa. The insect visits to flowers are correlated with the pollination-ecologically important floral character states as they are found in Hegi (1906-1931; 1963, etc.; 1966, etc.) and Knuth (1898a and 1899) by statistical analyses on an extensive data base. These analyses, based on a geographically restricted area, are correlated with the evolution of insect-feeding (and corresponding morphology) based on insect phylogeny and fossil record. These correlations require two presuppositions: “horizontal” uniformitarianism (Recent insect behaviour and physiology in taxa of lower rank are supposed to be rather constant over the world); and ”vertical” uniformitarianism (as it is used in historical geology). The results of these correlations form a series of functionally and stratigraphically/phylogenetically directed floral transformation series, dated by the fossil record of insects. The transformation series are compared with the fossil record of extant Angiospermae (for only a very minor part consisting of fossil flowers), requiring the possibility of spiral reasoning by the process of reciprocal illumination. The global scheme of the way of reasoning is illustrated in fig. 2-1. It is argued that the approach to describe evolutionary developments of pollination starting from insect fossil history and phylogeny is allowed, because in the course of evolution of insect-flower relationships, entomophilous flowers are more dependent on insect character states for (cross-)pollination, than insects are on floral character states for feeding. Free-living insects form an essential part of the environment of the “sessile” entomophilous flowers. New developments in entomophilous floral morphology are entirely dependent on the presence of corresponding insect character states. Only in case of success the, from that moment on more restricted (specialized), insect-flower relationships give rise to genuine co-evolution of anthophilous insects and entomophilous flowers. In chapter 3 the material and statistical methods are described, in addition to some prospects of this study in building a more refined survey, up to an analysis on the basis of extensive pollen analyses on the loads of the integuments and contents of the digestive tracts of central European Cetoniinae, Lepturinae and Cerambycinae of which the results are presented in the appendix. In chapter 4 a survey is given of the feeding-habits of the adult insects included in the present study as mentioned in Knuth (1898a and 1899) and of their relatives, giving a systematically complete survey of the feeding-habits of Coleoptera, Hymenoptera, Lepidoptera and Diptera (with some remarks on anthophilous insects of other orders), with stress on the taxa in which anthophily developed. The main conclusion of this survey can be that of many families the feeding-habits were not found in the more general entomological literature, and also in more specialized literature the feeding-habits in many cases are imperfectly known. Chapter 5 deals with the correlation of the anthophilous feeding-habits of insects with their phylogeny and fossil record. The correlations are restricted to the insect orders in which the main anthophilous insects occur: Coleoptera, Hymenoptera, Lepidoptera and Diptera. Comparison of the monophyletic Holometabola with their sister-group the Paraneoptera indicates that the original feeding-habits of both larvae and adult insects of the former were saprophagous and/or fungivorous in moist vegetable debris and similar substances (as already suggested by Tillyard, 1926). In describing the possible sister-group relationships within the Holometabola a non-pollination ecological conclusion is made: it is concluded that, if feeding on the blood of warm-blooded animals can be considered a synapomorphy in the Siphonaptera (fleas), they could not have originated earlier than in the Triassic (up to now, presumed origin of the Mammalia and Aves, see Romer, 1966); this is in contrast to Hennig (1981) who placed the origin of this order in the late Permian. In discussing the correlations of the feeding-habits of the taxa of lower rank with their fossil record and phylogeny within the orders mentioned above, the following conclusions are drawn. Coleoptera The habitat of the earliest Coleoptera most probably was that under the (loose) bark of trees. This habitat formed the selective pressure which favoured the adaptive development of elytra from the forewings. The earliest Coleoptera (Ommadidae) may have fed on reproductive structures or flowers as early as the lower Permian. In the upper Triassic this type of feeding must have been established, because all polyphagan series were present in that time (may be only some are of lower Jurassic origin). During the Jurassic the Polyphaga in which anthophily developed radiated considerably and in the upper Jurassic they were fairly well-differentiated (see tables 5-1 and 7-1). During the Cretaceous the differentiation of these beetles had a much slower rate, but, regarding the rather extensive differentiation in the upper Eocene to lower/middle Oligocene, at the end of the Cretaceous the origin may have laid of many (stem-groups) of families. The following reconstruction of relationships between anthophilous Coleoptera and reproductive structures and/or flowers can be made. After a fairly rapid radiation during the upper Triassic and Jurassic, in the Cretaceous only few developments took place, but correlated with the ecological break-through of the Angiospermae in the Upper Cretaceous, the differentiation rate increased again. Because it is not known when the origin of the Angiospermae has to be placed in the geological time scale (see below), in table 7-1 the stratigraphical appearance of extant insect taxa in which anthophily developed is given from the upper Triassic to the Lower Cretaceous. Hymenoptera The original feeding-habits of the Hymenoptera most probably were fungivorous. The habitat originally may have been decaying vegetable debris, but the Symphyta s.str. most probably lived in crevices of bark and under (loose) bark. The latter habitat probably formed the selective pressure which favoured the adaptive development of the cenchri, providing the possibility of fixing the wings in crawling under bark. The evolutionary developments towards and within hymenopteran anthophily were correlated with those in the maternal care and social behaviour. The earliest Hymenoptera (Xyelidae) may have fed on reproductive structures and/or flowers as early as the Triassic. The extant families of the Symphyta s.str. in which anthophily developed, differentiated from about the lower/middle Jurassic onwards and were definitely established in the upper Jurassic (see tables 5-2 and 7-1). In the Lower Cretaceous the earliest Apocrita appeared (Ichneumonoidea) and in the Upper Cretaceous an extensive differentiation of wasp families, in which, mainly nectarivorous, anthophily developed, can be observed (table 5-2). The differentiation of the Apoidea was about completed in the upper Eocene to lower/middle Oligocene and this may mean that the earliest Apoidea or their direct predecessors were already present in the uppermost Cretaceous. The following reconstruction of relationships between Hymenoptera and reproductive structures and/or flowers can be made. After a fairly rapid differentiation in the Jurassic of the Symphyta s.str., only some differentiations in the Apocrita, in which anthophily developed, took place in the Lower Cretaceous, followed by an increased differentiation after the ecological breakthrough of the Angiospermae and of the nectarivorous (aculeatan) wasps in the Upper Cretaceous. Probably in the same time the Apoidea originated and differentiated very rapidly during the (uppermost Cretaceous and) early Tertiary. The stem-group of the Bombinae and Euglossinae must have been present in the upper Eocene to lower/middle Oligocene. In correlating the feeding-habits with the phylogeny and fossil record of the Aculeata s.str. a pollination-ecologically interesting character state was met in the Bradynobaenidae. Although no feeding-habits were found for this family, the pinnate setae of its subfamily Typhoctinae may indicate close relationships with flowers (parallel evolution of setae as in Apidae). The morphology of the Bradynobaenidae urgently requires further research into their feeding-habits and behaviour. Lepidoptera When the Lepidoptera definitely originated within the Amphiesmenoptera (Lepidoptera + Trichoptera) is not known, but the earliest, mandibulate, certain Lepidoptera were present in the lower Cretaceous (Micropterigidae) and probably their predecessors were already present in the upper Jurassic. The Micropterigidae most probably fed on pollen in those times. The Ditrysia were very differentiated in the middle Tertiary. This differentiation (see table 5-3), particularly that of the Papilionoidea, indicates a much earlier origin. It is suggested that after the mandibulate, pollen-feeding phase, the development of a longer haustellum induced a rapid differentiation of the higher Lepidoptera after the ecological break-through of the Angiospermae in the Upper Cretaceous, and mainly took place in the uppermost Cretaceous and early Tertiary. Diptera The earliest probably pollen-feeding Diptera are found in the Bibionomorpha which were present in the upper Triassic (whether Tipulomorpha, Psychodomorpha and Culicomorpha also had relationships with reproductive structures and/or flowers in that time, or only developed nectarivorous anthophily from the Upper Cretaceous onwards, is not known). A rather rapid differentiation of Diptera, in which anthophily developed, can be traced in the Jurassic (see tables 5-4 and 7-1). During the Cretaceous new differentiations in dipteran anthophily hardly took place, but the extensive established differentiation of Brachycera in the upper Eocene to lower/middle Oligocene indicates a much earlier origin. It can be suggested that after the ecological break-through of the Angiospermae the anthophilous, mainly nectarivorous Brachycera differentiated rapidly in the uppermost Cretaceous and first half of the Tertiary. Chapter 5 is concluded with a survey of the stratigraphical appearance of the insect taxa in which anthophily developed. In chapter 6 the statistical analyses of the central European flower visits of insects and their association with floral character states are carried out. The differentiated survey of the correlation of the feeding-habits with the phylogeny and fossil record is narrowed to insect taxa of higher rank: Coleoptera, Diptera, Lepidoptera and Hymenoptera (section 6.1). The last order is divided into two groups, viz. the Apoidea and non-apoid Hymenoptera on the basis of their different morphology and (in many cases) behaviour (Apoidea being nearly obligatory anthophilous). Generalized special morphological character states are indicated for the insect groups (lengths of mouth parts). The interdependence of insects and floral character states is correlated with the results of chapter 5 (correlation of feeding-habits with the phylogeny and fossil record of the insect taxa in which anthophily developed). With regard to the complex of facultative and obligatory pollination types (section 6.2), it is concluded that the more specialized insect-flower relationships developed during the latest Cretaceous and became finally established in the middle Tertiary (possibly with the exclusion of more specialized pollination by beetles, which may have evolved earlier). The earliest Angiospermae, then, most probably were non-specialized entomogamous, depending at their time of origin, on pollination by Coleoptera, Diptera and non-apoid Hymenoptera (Symphyta). Comparison of the frequencies of obligatory pollination types within the complex of facultative and obligatory ones, indicates that bees and butterflies reached the highest degree of specialization of insect-flower relationships as far as it concerns the central European area (section 6.2.2). With regard to the obligatory pollination (section 6.2.2), corresponding with the pollination syndromes, it can be stated that obligatory myiophilous flowers (not sapromyiophilous) with some depth effect may have existed from about the middle Cretaceous onwards and gave rise to melittophilous and psychophilous flowers in the (uppermost Cretaceous and) early Tertiary, while at the same time the obligatory form disappeared. Phalaenophilous flowers developed later in the Tertiary. In the blossom-pollinator relationships (section 6.3) the following can be suggested. Allophilic flowers occurred in the earliest Angiospermae. Hemiphilic flowers may have originated in the late Cretaceous, giving rise to euphilic flowers in the early Tertiary. As regards the development of flower types (section 6.4) it can be stated that actino- or already pleomorphic flowers occurred in the early Cretaceous. From about the middle Cretaceous somewhat stereomorphic flowers, particularly with regard to pollination by longer-rostrate flies, developed, and may have given rise to zygomorphic flowers. Stronger zygomorphic and stereomorphic flowers became functional during the early and middle Tertiary. Among the earliest Angiospermae the following blossom classes (section 6.5) may have been present: inconspicuous entomophilous blossoms, dish- to bowlshaped blossoms and possibly some type of brush-shaped blossoms (the last ones not specialized for longer-tongued insects). From about the middle Cretaceous developments towards bell- and funnel-shaped blossoms were possible. In the late Cretaceous possibly more trumpet- and tube-shaped blossoms may have become functional and developed during the early Tertiary. Gullet-shaped blossoms are of later Tertiary origin. Flag-shaped blossoms are known from the late Paleocene. As regards the single flower as a pollination unit versus (small-flowered) inflorescences (section 6.6) it is argued that the most plesiomorphous entomogamous Angiospermae had single flowers as pollination units (considering only the size of the flowers with regard to the size of their potential pollinators; the flowers were arranged in inflorescences, see section 7.1) and the development of small-flowered inflorescences took place under the selective pressure of ovule damage by larger, mandibulate, partly anthophagous pollinators, through plants or large inflorescences with small flowers, dependent on pollination by minute insects. The more compact inflorescences are adapted to pollination by larger and stronger-flying pollinators with the advantage of some spreading of the small one- or few-ovuled ovaries. The development of small-flowered plants and inflorescences could have taken place very early in the evolution of the angiosperm flowers. The developed or rudimentary perianth (section 6.7) cannot be correlated to certain insect groups. Absence of perianth parts is mainly characteristic for anemophilous flowers. With regard to floral colours (sections 6.8.1 and 2) a comparison is also made between the taxa of lower rank by reciprocal averaging. In the comparisons the colour vision of insects is also included. One of the results of the analyses is, that in all insect groups the most plesiomorphous, anthophilous representatives mainly visit yellow (and white) flowers. The most plesiomorphous angiosperm floral colour most probably indeed was yellow (this may be supported by floral biochemistry). The total result consists of the following transformation series: —from (green) yellow to white: very early in angiosperm evolution; —from white to blue and blue-mixed colours: starting in the late Cretaceous; —from yellow or white to red and red-mixed colours: starting in the late Cretaceous; —from blue or blue-mixed to red and red-mixed colours and vice versa: from the late Cretaceous onwards. It generally can be stated that floral colours in the course of angiosperm evolution fanned out from the middle of the spectrum towards the extremes; towards the short wave-lengths in connection with pollination by higher Apoidea and Lepidoptera; towards the longer wave-lengths in connection with pollination by certain taxa of Lepidoptera and, later, birds. Cryptantherous flowers (section 6.9) may have been present from the late Cretaceous onwards, perhaps sapro-entomophilous inflorescences excluded. In the numbers of stamens per flower (section 6.10) the following developments could have taken place. The earliest Angiospermae most probably had many stamens. In conjunction with the developments towards smallflowered plants and inflorescences, reduction of the numbers may have occured very early in the evolution of the angiosperm flowers. Considering that obligatory anemophilous flowers have only very few stamens per flower, this can be regarded as an apomorphous condition. In larger flowers in the second half of the Cretaceous the stamens could become reduced in numbers, in connection with specialization of the pollination. With regard to the position of the ovaries (section 6.11), it can be stated that the selective pressure (mandibulate, anthophagous insect visitors, mainly beetles) favouring the development of half- to entirely inferior ovaries existed from the very origin of the Angiospermae and was probably completed very early. In the number of ovules per stigmatic surface (mostly corresponding with the carpel) (section 6.12), the number of ovules increases with an increase of the specialization of the pollination. The earliest entomogamous Angiospermae probably had few ovules per carpel. Reduction to one took place in the development to small-flowered inflorescences and anemogamy. Increase in specialized pollination types probably started in the Upper Cretaceous and early Tertiary in connection with specialized longer-tongued pollinators. As regards nectar presence and position (section 6.13) it appeared that it cannot be suggested on the basis of the statistical analysis whether the earliest Angiospermae had pollen or nectar-containing flowers. It is demonstrated that beetles in many cases (also) feed on pollen on nectar-containing flowers (see also the tables in the appendix). In the position of the nectar the following transformation series exists: from free by way of half-concealed to entirely concealed nectar. In the variation in time between receptivity and dehiscence (section 6.14.1) it is demonstrated that anemophilous flowers most often are protogynous, although this is also common among entomophilous flowers. Facultative and obligatory homogamy is mostly found in the obligatory pollination types. Protandry appears to be almost exclusively restricted to entomophilous flowers, but it requires adjacent developments to guide visiting insects over the flower or inflorescence by colours and/or odour and nectar, to optimalize the chance of cross-pollination. It is argued that a gymnospermous sister group of the Angiospermae most probably was protandrous, because of the fact that, in case of endosperm proliferation before fertilization, the simultaneous presence of pollen and endosperm in hermaphroditic, entomophilous reproductive structures has to be minimalized to avoid insect injury to the endosperm. In the case of homogamy and protogyny too much endosperm would be present simultaneously with the pollen, and mandibulate insect visitors (beetles and possibly also symphytan wasps) would easily feed both on pollen and proteinrich endosperm. In order to avoid endosperm injury, the simultaneous presence of endosperm and pollen must have included a minimum quantity of the first and have lasted as short as possible. This circumstance can only occur in protandrous conditions. Moreover protandry representing the most logical sequence of ripening of the end of a stem, the simultaneous presence of pollen and endosperm, given the insects present in those times, must have prohibited the development of absolute homogamy and protogyny (avoiding too much “advertized” endosperm in entomophilous, hermaphroditic conditions). Because protandry, however, requires adjacent “guiding”-developments, which obviously did not take place in the Gymnospermae, a selective pressure on the hermaphroditic reproductive structures, favouring cross-pollination, favoured the development of unisexual flowers (with a good chance to become anemogamous), and another selective pressure led to the development of “double fertilization” in which the endosperm only starts proliferation after fertilization. Once established, the “double fertilization” opened the way for the development of homogamy and protogyny. The formation of stigmatic surfaces can be considered an adaptation to avoid fertilization-decreasing pollen droplet-feeding by, also non-mandibulate, insects (flies). Analysis of the protandry in several families of the central European area, particularly in the Apiaceae, indicated that this type of dichogamy easily leads to the development of unisexual flowers, as so to avoid too much selfpollination. Protandrous flowers more often are self-incompatible than are protogynous ones. Protandry appears to be more characteristic for unspecialized entomophilous pollination types, than for specialized ones (creating second order herkogamy in the former; in the more specialized flowers, particularly in melittophilous ones herkogamy is present morphologically). The presupposition of this section was that unspecialized, mandibulate, injurious insect visitors may have formed the selective pressure favouring the development of second order herkogamy. The analysis does not support this. Dichogamy is developed under the selective pressure of optimalizing crosspollination, in protandry easily leading to unisexual flowers. Absolute homogamy (section 6.14.2) more often occurs in specialized insectflower relationships than in unspecialized ones. In the latter case it may induce the development of unisexual flowers (but to a lesser degree than does protandry). In more specialized pollination types it may have induced development of heterostyly, and under unfavourable pollination conditions it easily led to cleistogamy. In the developments towards dicliny (section 6.15) the following difference between anemogamous and entomogamous species is hypothesized. The difference of directability of the pollen vectors led to an enlarged chance of development of monoecy under protandrous/homogamous, anemogamous conditions and of dioecy under similar entomogamous conditions. The abundance of hermaphroditic protogynous flowers in anemogamy requires extension of the definition of the syndrome of anemophily: “unisexual flowers, in case of monoecy protogynous, or protogynous hermaphroditic flowers”, in stead of solely “unisexual flowers”. In section 6.16 the preceding evolutionary developments are outlined in a series of functionally and stratigraphically/phylogenetically directed transformation series. Chapter 7 deals with the reconstruction of the evolutionary developments in pollination ecology. In hypothesizing the theoretical model of the earliest angiosperm flowers (section 7.1) by considering the plesiomorphous states of many of the transformation series, added with some comparative-morphologically derived ones, the following conclusions were drawn. The earliest angiosperms most probably had the following character states. The pollen flowers were arranged in inflorescences, most probably leafy cymes, and were of intermediate size. They were hermaphroditic, protandrous and allophilic, non-specialized entomophilous, pollinated by beetles, symphytan wasps and oligoneuran (possibly mainly bibionomorphan) and possibly asiliform flies. Their general appearance was dish- to bowl-shaped, paleomorphic. The optical attractant was formed by the stamens and discoloured upper leaves (to some degree already perianth-like); the floral colours most probably were yellow and white. The many stamens were arranged in bundles which represented compound sporangiophores, bundles being formed in a phyllotactic spiral, continuing that of the floral envelopes. The stamen filaments were short. The gynoecium consisted of probably many superior, petiole-like stiped, at anthesis not entirely closed, carpels, which were infolded cupules with sessile stigmas. Each carpel contained a few ovules in a laminar or submarginal position. Double fertilization was present. The basic synapomorphy of the Angiospermae is the double fertilization and it strongly indicates an entomophilous, hermaphroditic origin. The indication of a selective pressure favouring the development of double fertilization, may favour the possibility of a polyphyletic origin of the Angiospermae. But the restricted possibilities of hermaphroditism in their predecessors and the probable uniqueness of this development, may indicate monophyly. In the food supply for insect visitors it is hypothesized that in first instance the fertile parts (being the most attractive for insect visitors) provided food in the following transformation series: 1) from plain stigma to stigma with glandular warts; 2) from stamens to staminodes; 3) from plain stamens to stamens with glandular bodies at the base of the filament; 4) from plain stamens to stamens with glandular bodies at the apex of the connective; 5) from plain staminodes to staminodes with glandular bodies; 6) from stamens with glandular bodies to staminodes with the same; The staminodes with glandular bodies occasionally may have given rise to tepals and petals with nectaries. In increasing the attraction of the flowers, laminar stamens were derived from stamens with a slender filament and by reduction of the upper leaves. The laminar stamens gave rise to tepals by way of laminar staminodes. This means that there more probably is question of sterilization of laminar stamens, than of the reverse. In the remarks on the origin of the Angiospermae (section 7.2) it is concluded that there may be possibilities of temporary presence of hermaphroditic reproductive structures in the Bennettitales (demonstrated) and Gnetales (most probably). In the Pentoxylales these are not known. The development of double fertilization is functionally explainable, and thereby it is less probably a development according to the “economy principle”. The probable entomogamy of the earliest Angiospermae may include that the “upland” theory is superfluous, because entomophilous pollen is quantitatively rare in stratigraphical pollen analysis. In section 7.3 the apomorphies of the transformation series mentioned in section 6.16 are approximately dated on the basis of the fossil record and phylogeny of the insect taxa in which anthophily developed. Added is the dating of ornithophilous and chiropterophilous pollination types, based on the fossil record of birds and bats in which anthophily developed, respectively. In section 7.4 the fossil record of extant angiosperm taxa is correlated with the dated apomorphies in section 7.3. For each taxon an estimation is made of the pollination type, based on the presence of extant insect taxa in which anthophily developed in the period in which the angiosperm taxon appeared, and on the basis of the type of fossils (macrofossils or fossil pollen) and the abundance of the fossil pollen. The types of angiosperm fossils may form an indication of the pollination type of the flowers they represent. There are three possibilities: 1) if in a period macrofossils are present and fossil pollen of the same taxon is not found in stratigraphical pollen analysis, or does not occur in some abundance (pollen not originating from the macrofossils), there is a major chance that the fossils represent entomophilous (or otherwise zoophilous) flowers; 2) if in a period both macrofossils and in considerable quantities fossil pollen of the same taxon are found (pollen not originating from the macrofossils), there is a major chance that the fossils represent anemophilous flowers; 3) if only stratigraphical pollen is found in considerable quantities, it most probably represents anemophilous flowers. Analysing the survey of fossils of extant angiosperm taxa at the end of this section, it is concluded that fossil pollen representing entomophilous flowers in quantitative, stratigraphical pollen analysis is rare and occurs irregularly in time (qualitatively, however, more entomophilous than anemophilous pollen types are found). Macrofossils qualitatively more often represent entomophilous than anemophilous flowers. Pollen morphology also can give an indication of the pollination type: sculptured, tectate (and sticky) pollen mostly is entomophilous and psilate, and atectate (dry) pollen mostly is anemophilous. With regard to the pollination type, it is suggested that entomophilous pollen is longer on its way from locule to stigma, than is anemophilous pollen, and therefore more urgently needs a protective layer of “pollenkitt” to avoid desiccation; at the same time it functions as adhering agent to the body of the pollen vector. It is suggested that the columellate sculpture of the exine keeps the protective “pollenkitt” equally distributed over the pollen surface, and that the combination of sculpture and “pollenkitt” has a function in itself in the relation of adhering to the body of the pollen vector and the deposition on the stigmatic surface. Because recent investigations indicate the presence of columellate pollen in as early as the Triassic and because also in gymnosperms evidence for relicts of columellae is found, and the earliest angiosperms most probably were entomogamous, it is suggested that the earliest angiosperm flowers had columellate pollen, which would mean that atectate, psilate pollen (as e.g. in the Degeneriaceae) has to be regarded as apomorphous. The correlation of the extant angiosperm taxa with the above mentioned elements indicates the following appearance of floral character states in the stratigraphical periods: Barremian-Albian —anemophilous flowers, including possibly already inconspicuous, green flowers with reduced perianth, few stamens and a reduced number of ovules per stigma (carpel); the presence of this apomorphous pollination type and its corresponding probable character states indicate a much earlier origin of the Angiospermae; —some more obligatory beetle pollination; —possible sapro-entomogamy; —somewhat hemiphilic flowers are possible (in connection with more obligatory beetle pollination); —possible haplomorphic flowers; —actino- and pleomorphic flowers; —probable (partial) sympetaly; —dish- to bowl-shaped blossoms; —brush-shaped blossoms (not specialized for longer-tongued insects); —small-flowered inflorescences; —yellow and white floral colours; —reduction of the numbers of stamens per flower in connection with smallflowered entomophilous inflorescences; —reduction of numbers of ovules per stigma (carpel) in connection with smallflowered inflorescences; —possible protogyny; —possible unisexual flowers. Cenomanian-Turonian —early more obligatory myiogamy; —more obligatory cantharogamy; —early more obligatory wasp pollination; —earliest somewhat stereomorphic and possibly zygomorphic flowers (e.g. margin flowers of small-flowered inflorescences); —earliest possible bell- to funnel-shaped blossoms; —larger solitary flowers; —possibly some blue and blue-mixed colours in connection with early more obligatory myiogamy; —peri- and epigyny; —nectar-containing flowers with open nectar. Coniacian-Santonian-Campanian —possible very early pollination by Apoidea or their