Trees (1995) 9:269-278

9 Springer-Verlag 1995

The significance of structure for imbibition in seeds of the Norway spruce, Picea abies (L.) Karst. Eila Tillman-Sutela 1, Anneli Kauppi2 1The Finnish Forest Research Institute, P. O. Box 16, SF-96301 Rovaniemi, Finland 2 Department of Botany, University of Oulu, SF-90570 Oulu, Finland Received: 10 May 1994/Accepted: 3 November 1994

Abstract. Since the observations of those regularly handling Norway spruce [Picea abies (L.) Karst.] seeds with regard to their imbibition frequently disagree with earlier opinions that this process is markedly inhibited by the seed coat, we decided to examine the morphological factors influencing imbibition in seeds of different colour and different provenances. The seed coat, consisting of the sarcotesta, sclerotesta and endotesta, was found to have little influence on the passage of water, despite the presence of sclereids full of wax lamellae. No differences in seed coat structure were observed between provenances or colours of seeds. The cells of the endotesta were lignified in the area of the micropyle, however, and stood out lip-like on the outer surface of the micropyle after imbibition. An opening in the sclerotesta filled with parenchyma cells was also seen at the chalazal end of the seed. Neither of these openings, which were covered by accumulations of wax, served as the main route for the passage of water, though the micropyle opened up slightly after only 24 h incubation, when the lignified cells bordering it swelled differently from the rest of the endotesta. The progress of water into the seed soon discontinued, however, as the tip of the nucellar cap, covered with wax and crystals, effectively plugged the micropyle. This opening of the micropyle may be the reason why the IDS method does not always succeed in separating viable from non-viable spruce seeds sufficiently well by their density. Imbibition was mostly regulated by the lipophilic layers surrounding the endosperm, which are mainly of nucellar origin, and particularly the megaspore membranes, the outer and inner exine. Imbibition was further hampered by the impermeable nucellar cap, which covered about 3/4 of the length of the endosperm and had merged with the outer exine at its edges. Deposits of wax were observed both between the exines and between the endotesta and the nucellar layers at the edges of the nucellar cap. Waxes may serve as a defence

Correspondence to: E. Tillman-Sutela

against diseases at the sites of water penetration, while simultaneously increasing the significance of the nucellar endosperm covers as regulators of imbibition.

Key words: Picea abies (L.) Karst. Morphology - Imbibition - Permeability

Seed coat -

Introduction Research into the imbibition and germination of conifer seeds has often treated pine and spruce as such closely related genera that results obtained for one apply to the other. However, most observations of this kind have been derived from germination tests (Goo 1951; Zentsch 1962; Hagner 1985; Bergsten 1987). No thorough attempt has been made to determine the morphology of the seed coat or the layers inside it, especially since macroscopic examination suggests that pine and spruce seeds have a broadly similar anatomy in this respect (Nobbe 1876; Kujala 1927; Zentsch 1960, 1962; Sarvas 1964; Hagner 1985). Anatomical descriptions of seeds of the Pinaceae have been produced only by light microscopy, and mostly for the genus Pinus (Ferguson 1904; Coulter and Chamberlain 1910; Schnarf 1933, 1937; HLkansson 1956; Sarvas 1962; Voroshilova 1983). Moreover, since most of the research has been concerned with the ovule or the early developmental stages of the seed, information on the structures of the mature spruce seed coat and the endosperm covers inside this has remained incomplete and the terminology as vague as that for the pine seed. The terminological clarifications and the origins of the anatomical parts and cell layers have been presented in our earlier paper (Tillman-Sutela and Kauppi 1995, Table 1). For instance, when we use the term "endosperm" in this text, we are referring to the primary endosperm of the female gametophyte typical of gymnosperm seeds. We have also used the term "megaspore membranes", which is generally used in

270 literature (Schnarf 1937; von Ltirzer 1956; Singh 1978) when referring to the layers closest to the primary endosperm. Similarly, the treatment instructions to be used with the IDS method (I = incubation, D = desiccation, S = separation) for sorting viable from non-viable seeds were developed primarily for pine seeds, although they have also been applied to spruce (Bergsten 1987). Practical observations on the differences in the imbibition behaviour of seeds of N o r w a y spruce have been made, especially in Finland, where the IDS method has been used for forestry purposes for several years. Spruce seeds frequently begin to germinate even during incubation, or else it may prove impossible to achieve a sufficient moisture difference to separate viable from non-viable seeds (A. Varjo, personal communication). Similar problems have been encountered with the sedimentation o f white spruce seeds, Picea glauca (Moench) Voss, by the IDS method (Downie and Wang 1992). The difference in the rate of imbibition between the seeds of N o r w a y spruce, Picea abies (L.) Karst., and Scots pine, Pinus sylvestris L., have been attributed to the fact that the nucellar cap covering the micropylar end of the endosperm is only partially permeable and is larger in the spruce seed (Zentsch 1960, 1962). More recently, however, it has been found that the nucellar cap in Scots pine seeds is i m p e r m e a b l e and only the endosperm covers are partly permeable. The seed coat itself has proved to be o f little significance as far as the differences in imbibition between pine seeds o f different provenance are concerned, although it has morphological differences related to provenance and the colour o f the seeds (Tillman-Sutela and Kauppi 1995). The authors know of no existing documented accounts of the significance of the structure o f the seed coat and the layers inside it for the imbibition of seeds of Picea abies. The purpose of the present paper is to explore the structure of these parts of the spruce seed and to assess their role in imbibition and with respect to possible differences between seeds of different colour and provenance.

Materials and methods The material used for the experiments consisted of commercial seeds of Norway spruce, Picea abies (L.) Karst., of three provenances: Sodankyl~i (67~ Rovaniemi (66~ and Korpilahti (62~ The seeds had ripened in 1989, and had been stored at a temperature of +2 ~ Their moisture content was about 6%. A 50 g batch of seeds of each provenance was taken in accordance with the International Seed Testing Association (ISTA 1985) instructions and classified by eye into three colour categories: dark, light and mixed. Only the dark and light seeds were used in the experiments. The surface structures of the seed coat, the endosperm covers and the endosperm of dark and light seeds of all three provenances were examined with a JMS 6300F field emission scanning electron microscope. The seeds were soaked for 24 h at 20 ~ cut either longitudinally or transversely, and the nucellar cap, the endosperm and the layers covering it were examined separately. The samples were fixed in FAA (ethanol, glacial acetic acid and formalin 85:5: 10), dehydrated in an alcohol gradient, critical-point-dried, attached to a SEM mount, and sputtered with gold. The wing structure and attachment were observed using storage-dry seeds, mounted and sputtered with gold.

Samples for light microscopy (Leitz-Dialux 20 EB) were obtained by soaking seeds in water for 48 h, splitting them and fixing the halves in FAA, dehydrating them in an alcohol gradient, infiltrating them with plastic for 48 h at +20 ~ and embedding them in Reichert-JungHistoresin (70-2218.500). A total of approximately 720 sections about 4 ~tm thick were produced with a glass knife on a LKB-Historange microtome, which were either longitudinal sections or cross-sections from the micropylar and chalazal end. Some of the preparations were stained with methylene blue azure (Humphrey and Pittman 1974). Histological staining was performed with aniline sulphate for the identification of lignin-containing tissues, Sudan Black B for lipophilic tissues and ruthenium red for pectinous tissues (Jensen 1962). Both stained and unstained sections were also examined under a fluorescence microscope with UV excitation (Nikon Optiphot 2 with a 100 W HG lamp and an automatic Nikon Microflex UFX 35 DX camera, filter combination Ex 365/10, DM 400, BA 400). Changes in the nucellar cap covering the micropylar end of the endosperm in the course of germination were observed by ruthenium red staining. The permeability of the seed coat was analyzed by placing 150 light and 150 dark seeds of the northernmost (Sodankyl~) and the southernmost (Korpilahti) provenances in a methylene blue solution (Merck: E.9. No.52015;S.No.1038). Four seeds from each batch were then split and the structural parts were removed for examination under a stereo microscope first at 3 h intervals and later at 12 h intervals for a period of 8 days.

Results Surface structure and basic anatomy The seeds were seen clearly to be composed o f two halves, with curved ridges at their junction, when examined by SEM (Fig. 1A). Where the wing was still present, it covered the upper surface of the seed and was attached to the edges of the ridges (Fig. 1 B). The halves were covered by the sarcotesta, the outermost layer o f the seed coat, formed of parenchyma cells, which had holes and breaks in it at some points on the upper surface of the seed and at the edges of the ridges, from which the wing had detached itself (Fig. 1 C). The sarcotesta on the lower surface of the seed was largely unbroken and covered by a waxy layer (Fig. 1 D), which was thickest at the chalazal and micropylar ends of the seed (Fig. 4 A, G). The light microscopy sections showed the sarcotesta to be thin in all the provenances studied, being frequently composed of only one cell layer (Fig. 4 C). In a macroscopic examination, the sarcotesta cells were reddish brown in the dark seeds, but scarcely pigmented at all in the light seeds. The main part of the seed coat consisted of the sclerotesta, which consisted of lignified isodiametric brachysclereids (Figs. 2 B - C , 4 C - D ) . The sclerotesta was thickest at the micropylar and chalazal ends (Figs. 2 A , 4G). All the provenances had 3 - 5 layers o f sclereids of different orientation, with a particularly obvious change in orientation as they approached the micropyle. The lignified parts of the cell walls were fairly thin and full of small, simple pits throughout (Figs. 2 B, 4 C - D ) . The intercellular spaces in the sclerotesta formed narrow canals (Fig. 2B). The sclereids were filled with wax lamellae (Fig. 2 C - D ) . Only the thin parts of the cell walls in the sclerotesta fluoresced in UV light (Fig. 4 E - F ) and were stained yellow by aniline sulphate.

271

Fig. 1 A - F . Structure of the seed coat of Picea abies (L.) Karst. A Surface of the seed at the micropylar end, showing the ridge (H) joining the two halves, the sarcotesta of the lower surface (SAD and the sarcotesta of the upper surface (SA~). B A winged seed (W) viewed from the lower surface (SAD, with micropyle (M) and chalaza (CHA). C Magnified view of the sarcotesta (SA) on both sides of the ridge (H). The sarcotesta of the upper surface (SA2) has holes (white spots) in it.

I) Magnified view of the unbroken sarcotesta on the lower surface of the seed (SAI). E Endotesta (EN) protruding from the micropyle opening in a lip-like manner. F Endotesta (EN) at the edge of the micropyle (M), shown in a longitudinal section of the seed. Note the change in the endotesta at the point marked with an asterisk, also showing the sarcotesta (SA) and sclerotesta (SCL)

272

273

An opening filled with parenchyma cells and covered by a thick layer of wax was observed in the otherwise dense sclerotesta at the chalazal end of the seed, with some longitudinal tracheary elements in its central part (Fig. 4 A - B ) . The cells located at the ridge joining the two parts of the seed also stood out from the surrounding sclereids as having thicker walls, but there was no opening in the sclerotesta. A tissue which protruded lip-like at the micropyle was distinguishable on the surface of a moistened seed (Fig. 1 E). In a longitudinal section, this tissue proved to be the endotesta, the innermost layer of the seed coat (Fig. 1F). The endotesta was made up of parenchyma cells, often flattened in shape (Fig. 4C), which could also be distinguished from the other layers of the seed coat in UV light, as a brown, non-fluorescent layer (Fig. 4 F). Only at the opening of the micropyle was it lignified and autofluorescent under UV light (Fig. 4 H). A number of thin layers of collapsed nucellar cell walls, located beneath the endotesta and attached to it in places (Fig. 2 F), surrounded the endosperm. These nucellar layers stood out at the chalazal end of the seed, still forming a tissue which was attached to the seed coat over a long distance (Fig. 4 A). This partly degenerated nucellar tissue was found all over the unripe seeds (Fig. 4C). A nonfluorescent mass, which stained yellowish with methylene blue, was observed between the endotesta and these endosperm covers in places (Figs. 2E, 4E). In this case, a layer could be distinguished in the nucellar layers which appeared greenish upon methylene blue staining (Fig. 4 D) and was also shown to be lipophilic. The layers closest to the endosperm and the cuticle covering it consisted of the megaspore membranes, which could be distinguished from the nucellar remnants by their pronounced blue fluorescence under UV light (Fig. 4F) and their blue-green colour upon methylene blue staining (Fig. 4C). Digitiform baculae, about 4 ~tm in height, were observed on the surface of the inner exine (Fig. 2 F - H ) . These baculae were about 0.5 ~tm in diameter and, where the thin layer of outer exine covering the baculae had peeled away, they remained attached to the inner exine and the intine possibly fused with it (Fig. 2 E - F ) . The intine could not, however, be distin-

Fig. 2 A - H . Structural details of the seed coat and the layers inside it in the seed of P abies. A Longitudinal section of the micropylar end of the seed. Note the thickening of the seed coat (SC), the endosperm (E), the embryo (EM) and the nucellar cap (NC). B Sclereids with pits (P/) and narrow intercellular spaces (IS). C, D Magnified view of sclereids filled with wax lamellae (arrows). E Secretory material (asterisk) located between the endotesta (EN) and remnants of the nueellar wall layers (NR), also showing the sarcotesta (SA), sclerotesta (SCL) and megaspore membranes (arrows). Methylene blue staining. F Detached megaspore membranes inside the seed coat (SC), the nucellar wall remnants (NR), the outer (OEIO and inner exine (IEX) with digitiform baculae. The wax is shown by an asterisk. G Black-stained megaspore membranes (arrows) in a Sudan Black staining of a longitudinal section through the seed, nucellar remnants (NR), seed coat (SC). H Magnified view of the megaspore membranes (OEX and 1EX) and the digitiform baculae

guished in these samples. In the case of the seed preparations, these megaspore membrane layers remained on the endosperm surface at the chalazal end. At the micropylar end, however, they remained on the inner surface of the nucellar cap around the endosperm and were shown by Sudan Black staining to be lipophilic similar to the nucellar cap itself (Figs. 2G, 3E). The micropylar end of the endosperm was covered for about 2/3 of its length by the nucellar cap, which had longitudinal and transverse folds (Figs. 3 A - B ) . The cap was shrouded in a cuticle and was composed of numerous layers of overlapping, elongated parenchyma cells, and became thinner towards its edges (Fig. 3E). This highly UV-fluorescent cap was curled at its edges and was regularly fused with the other nucellar layers covering the endosperm and with the outer exine (Fig. 4 H - I ) . Wax was found between the endotesta and the nucellar layers close to this point of attachment. The tip of the nucellar cap was particularly heavily folded and its outer surface was likewise covered with a layer of wax (Fig. 3 D), and there was degenerated parenchyma inside it (Fig. 4G). The nucellar cap was so thin at its tip that some of the cells inside it became markedly elongated in the direction of the micropyle when the seed was soaked (Fig. 4 G - H ) , acting as a plug which closed the micropyle. The tip of the endosperm was covered by an affected tissue, which was strongly cutinized and had some secretory substance visible at both sides (Figs. 3 C, 4 G, 4 J). This substance stained blue with methylene blue, whereas the mucous cells adhering to the cap of the radicle stained violet (Fig. 4J). As the seed germinated, this substance was extruded out of the tear that had opened up in the nucellar cap as a sticky mass before the radicle and its cap emerged. Similar to the tip of the cap, it stained red with ruthenium red (Figs. 4 K - L ) . Circular agglomerations of crystals about 10 ~tm in diameter were observed on the surface of the megaspore membranes facing the endosperm at the chalazal end of the seed and near the edge of the nucellar cap (Fig. 3F). Separate tetragonal crystals were also encountered in the latter areas, being located inside the exine layers. Such crystals were detectable both in the micropylar opening and the degenerated parenchyma inside the tip of the cap and in the sclereids (Fig. 2D). Tetragonal crystals were seen at points of damage on the seed coat, where large numbers of microbes were also present (Fig. 3 G).

Permeability of the cell layers The stain penetrated the seed coat first close to the micropyle at the ridge joining the two halves of the seed (Fig. 1 A), where coloured patches were observed after only 3 h of soaking (Fig. 5 A). Somewhat later, after 6 h, similar patches were also seen at the chalazal end on the inner surface of the seed coat. It took about 9 h until the first coloured spots appeared on the outer surface of the nucellar remnants (Fig. 5 A), which were composed of several layers (Fig. 2 E - G ) . The tip of the nucellar cap covering the micropylar end of the endosperm (Fig. 3 A - B ) was also stained by that time (Fig. 5 A). The stain seemed to advance

274

Fig. 3 A - G . Nucellar cap and crystals in the seed of P. abies. A A peeled spruce seed with the nucellar cap (NC) and parts of the other nucellar remnants (NR) making up the endosperm covers. B Magnified view of the pronounced longitudinal and transverse folds on the tip of the nucellar cap (NC). C Secretory material (arrow) on the tip of the endosperm (E) inside the nucellar cap. D Wax-covered (arrow) tip of

the nucellar cap (NC). E Markedly lipophilic nucellar cap (NC) and the megaspore membranes inside it (OEX and IEX) in a Sudan Black staining of a longitudinal section, where the endosperm (E) is also visible. F Crystal agglomerations of the inner surface of the outer exine (OEX) at the edge of the nucellar cap. G Tetragonal crystals in an infected tear in the seed coat

275

Fig. 4. see legend page 276

276 equally rapidly through the seed coat in the seeds of both the northern and the southern provenance and in both the dark- and the light-coloured seeds. The stain penetrated the layers of the nucellar remnants inside the seed coat fastest at the area between the chalaza and the edge of the nucellar cap, where the first patches of colour on the surface of the endosperm were detected after 24 h (Fig. 5 B). The micropyle had also opened up distinctly in about 10% of the seeds by this time (Fig. 4 G - H ) , and the tip of the nucellar cap was stained a deep blue. The staining of the nucellar cap, however, did not seem to advance further through the micropyle opening. A uniform layer of wax was observed between the endosperm and the nucellar layers covering it. Again, the provenance and the colour of the seeds did not appear to affect either the penetration of the stain or the opening of the micropyle. The endosperm tissue was first affected by the stain in its central area and at the chalazal end, the part covered by the nucellar cap and the radicle of the embryo being the last to be stained (Figs. 2 A , 5 C). Only the edge and the area

Fig. 4 A - L . Morphology of the seed of P. abies. A Accumulation of wax (long arrow) at the chalazal end covering the opening in the sclerotesta (asterisk), the nucellar tissue (NR) inside it and the exine layers (short arrow) inside this. A longitudinal section stained with methylene blue, • 125. B Longitudinal tracheary elements (arrows) in the chalazal opening, stained red with methylene blue. • 250. C Three layers of pitted sclereids beneath a thin, partially pigmented sarcotesta (SA), a single endotesta layer (EN) and nucellar cells (NR) in an unripe seed. Note the digitiform baculae attached to the outer surface of the inner exine (arrow). Methylene blue staining of a cross-section. • 250. D Pitted sclereids and a distinctively stained cell layer (arrow) on the inside of these. Methylene blue staining of a longitudinal section. • 250. E Fluorescence microscope section of the sclerotesta and the layers inside it. Note the autofluorescent exine layers as well (arrow). • 250. F An unstained longitudinal section through the seed under UV light. The thin parts of cell walls in the sclerotesta fluoresce, but the wax lamellae, endotesta and nucellar layers, with the exception of the nucellar cap, stand out as brown in colour, i.e. non-fluorescent. The megaspore membranes (arrows) and the cuticle of the endosperm show blue fluorescence. • 125. G Methylene blue staining of a longitudinal section through the micropylar end of a moistened seed. Some of the tissue at the tip of the nucellar cap (NC) has become hypertrophied (asterisk) on absorbing water. The radicular cap is compressed and the endosperm tip tissue has stained reddish. The transition zone of the endotesta is indicated with arrows, x 60. H A fluorescence microscope view of the micropylar end of a moistened seed. The sharply folded nucellar cap stands out as white and the nucellar tissue inside it and the sclerotesta (SCL) are bluish. The endotesta is also autofluorescent (arrows) closer to the micropyle. An unstained longitudinal section. • 60. I An unstained longitudinal section through the centre of the seed under UV light, showing the attachment of the edge of the nucellar cap (NC) to the outer nucellar layers (black spot) covering the endosperm. The megaspore membranes (arrows) stand out clearly on the inner surface of the nucellar cap, but at the chalazal end the inner exine with baculae remains on the surface of the endosperm. The outer exine is seen on the inner surface of the nucellar layers in places, • 125. J A magnified view of the affected tissue (arrow) of the tip of the endosperm with its covering cuticle. Note the bluish secretory material inside it (asterisk). Methylene blue staining of a longitudinal section. • 125. K A germinating seed, with slimy material stained with ruthenium red issuing from a tear close to the tip of the nucellar cap (arrow). L A bisected germinating seed, with the radicle (arrow) protruding from the micropyle. Ruthenium red staining

Fig. 5 A-C. Penetration of the staining solution in the structural parts of the Norway spruce (P. abies) seed. The shading indicates the places where the staining solution was first observed on the tissue layer in question, m = micropyle, sc = seed coat, h = ridge joining the two halves of the seed coat, nr = remnants of the nucellar wall layers, nc = nucellar cap, iex = inner exine with baculae attached to it, e = endosperm, em = embryo. A The patches of colour in the seed coat and the nucellar layers inside it after 3-9 h soaking. B The first observations of coloured spots on the surface of the endosperm after 24 h. The nucellar cap covering the micropylar end of the endosperm and the exine layers attached to it have been removed. The exine layers often remain on the surface the endosperm at the chalazal end, as shown on the left. C Staining solution in the tissues of the endosperm and embryo around the edge of the nucellar cap after 84 h. The part of the endosperm which was covered by the nucellar cap is unstained

around the tip of the nucellar cap were coloured on the surface by that time, while the inside was still unstained. The first seeds to be stained fight through were observed after about 5 days of incubation. More variation in permeability appeared to exist within the batches of seeds than between them, as all the lots still contained some seeds stained only in patches on the inner surface of the seed coat after 8 days of soaking. It is also worth noting that the tissue at the chalazal opening of the sclerotesta (Fig. 4 A - B ) still appeared unstained when the seed was otherwise thoroughly blue.

Discussion The seed coat has, up till now, been regarded as the essential factor restricting imbibition and germination in spruce seeds (Zentsch 1962). The interpretations given to the differences in imbibition observed between spruce and pine seeds in everyday practice have remained somewhat vague, as the knowledge of the structure of spruce seeds has been inadequate. The interpretations given in the literature have also varied as to what is included in the seed coat. The present morphological investigation showed that although the seeds of Picea abies and Pinus sylvestris consist of largely the same structural elements, there are differences between them in the structure of the testa and nucellar layers, which are of significance for imbibition. As in the pine seed, only some of the layers covering the endosperm originate from the integument of the ovule and can thus be classified as part of the seed coat senso stricto (cf. TillmanSutela and Kauppi 1995, Table 1). The seed coat proper was found to have little effect in restricting the passage of staining solution into the stored spruce seeds of our trials. Both the sarcotesta and the endotesta were often only one cell layer in thickness and their cuticle invisible, so that water also passed through them without difficulty. Schnarf (1937) reports these cells

277 to have no cuticle. Similarly, although fresh unstored seeds had a wax layer covering the sarcotesta on their lower surface, it was not an effective barrier to imbibition, as the sarcotesta of the upper surface was exposed after the detachment of the wing. For the same reason, small holes and tears had developed at the edges of the ridge joining the two halves of the seeds, which again allowed the stain to penetrate rapidly through the seed coat in these places. The cells of the sclerotesta were full of wax lamellae, but these did not seem to impair permeability, as the intercellular spaces formed narrow canals and the lignified parts of cell walls were thin and full of pits. No differences in seed coat structure or permeability to the stain were observed between the provenances and the colours of spruce seeds, as had been the case with pine seeds (Tillman-Sutela and Kauppi 1995). On the other hand, the abundance of wax lamellae in the sclereids and the wax layer on the lower surface of fresh seeds may provide one explanation for the fact that it is not easy to bring the moisture content of spruce seeds up to the required 30% at the initial moisture stage prior to IDS incubation. According to our results, a more efficient barrier than the seed coat proper to the passage of water was formed by the dense layers of nucellar remnants in the cavity between the seed coat and the megaspore membranes adjacent to the endosperm. These nucellar layers were still in the form of individual cells in places at the chalazal end of the seed and were attached to the endotesta, so that it was impossible to remove them intact from the seed coat, as in the case of pine seeds (Hoff 1987; Tillman-Sutela and Kauppi 1995). The megaspore membranes at the chalazal end were frequently attached firmly to the surface of the endosperm and, when the exine layers became separated, the capita of the digitiform baculae attached to the inner exine remained as patches reinforcing the surface of the outer exine. These observations proved the outer exine at least in some Pinaceae, i.e. in Picea abies and Pinus sylvestris, to be structurally more solid than it has been previously thought, as the capita of baculae were not free but fused with each other as in Cupressaceae (von Ltirzer 1956). This structure partly explains the capacity of the exine layers to regulate imbibition. Thomson (1905) compared the megaspore membrane with the exine-intine of pollen grains, and considered the two structures to be fundamentally alike in their morphology and chemical composition. According to him, the outer exine is suberized, while the inner exine and the intine contain pectocellulose and, consequently, can be regarded as selective. This point of view is also in agreement with the observations that the exine of pollen consists of sporopollenin (Fahn 1990). Our histological staining showed that the cuticle of the endosperm and the layers covering the endosperm were lipophilic, which, apart from their structure, also affected their permeability. Even with the structural differences found in these layers, the results reinforced our earlier observations on pine seeds (Tillman-Sutela and Kauppi 1995) except that there was no divergence at this stage of imbibition between the provenances of spruce seeds. According to our results, it was the micropylar end of the endosperm that was best protected against imbibition in the spruce seed, as it was covered for about three-quarters

of its length by a dense, multilayered nucellar cap which had a cuticle of its own and was fused with the nucellar remnants at its edges. The results suggest that, as in the pine seed, the nucellar cap of the spruce seed is also impermeable rather than semipermeable, as reported by Zentsch (1962). This conclusion was supported by the lipophilic nature of the nucellar cap as shown by histological stainings and its pronounced UV fluorescence. The stain solution penetrated into the endosperm much in the same way as in pine seeds, despite the differences in the structure and size of the nucellar cap (Tillman-Sutela and Kauppi 1995). These observations shed further light on the published results, which suggest that spruce seeds placed in damp sand with the chalazal end downwards germinate more quickly than those placed with the micropyle downwards (Zentsch 1962). The results proved the abundance of wax and crystals to be characteristic of Norway spruce seeds, which is another previously unreported difference compared with Scots pine seeds. The preparation of samples revealed a uniform sheathing of wax inside the nucellar layers covering the endosperm. In fixed and dehydrated samples, wax was found especially in the area to which the nucellar cap was attached. At the same points, a distinctively stained cell layer was also found on the inside of the endotesta, possibly representing tapetum tissue that had remained active (Schnarf 1933; Singh 1978). Crystal formations were also noted close to the points where the water first penetrated the endosperm covers. Since both crystals and wax could be found in infected ruptures in the seed coat and at the tip of the nucellar cap, these seemed to be connected with the seed's protective and defensive mechanisms. This view is also supported by the earlier observations on crystals occurring in the damaged surface tissue of spruce needles (Fink 1991; Back et al. 1993). The thickest layers of wax were observed on the seed coat in places where the sclerotesta was found to have structural openings. Apart from the micropyle, a similar opening, though plugged with tissue, was only found at the chalazal end, whereas in pine seeds there was another opening in the central part (Tillman-Sutela and Kauppi 1995). Neither of these openings in the spruce seed functioned as the main penetration route for external water, confirming our pine seed observations. The thick layers of wax over the openings may be another feature of the seed's protective structure, providing support for the thin, only weakly pigmented sarcotesta. In the case of the pine seed, the phenoliferous sarcotesta is known to afford protection against both diseases and rapid imbibition (Grzywacz and Rosochacka 1980; Tillman-Sutela and Kauppi 1995). The special structure of the micropyle in the spruce seed, which was first reported here, explained part of the differences in behaviour during imbibition compared with the pine seed. The endotesta cells in the micropyle of the spruce seed were lignified and curved outwards from the seed coat like lips. Upon imbibition, they expanded differently from the remaining endotesta. The wax covering then broke and the micropyle opened. The passage of water by this route appeared to be interrupted fairly soon, however, as the tip of the nucellar cap with its covering of

278 crystals and wax plugged the micropyle at a time when imbibition of the endosperm and embryo had proceeded sufficiently far to ensure undisturbed growth. According to our observations, the dense layers inside the hypertrophied tissue of the nucellar cap prevented the water from reaching the endosperm via this route. It may, nevertheless, be this structural way of the micropyle to open that causes the difficulty noted in achieving a sufficient difference in density between viable and non-viable spruce seeds by the IDS method (Bergsten 1987; Downie and Wang 1992). Imbibition may also be affected by the large amounts of wax present, the composition and temperature behaviour of which is not known at present. On the other hand, these waxes, which were found here, may also be one reason for the narrow optimal temperature range for the germination of spruce seeds (Kamra 1967; Bergsten 1987; Leinonen et al. 1993). Together with the rapid opening of the micropyle, the abundance of waxes may be the main reason for the poor predictability of the germination parameters of water-separated, redried spruce seeds (Bergsten 1987; Downie and Wang 1992). Acknowledgements. The authors wish to thank Prof. Sirkka Kupila-

Ahvenniemi for her comments, Dr. Matti Kauppi for his valuable help, Tuula Vuorinen, for producing the drawings, Helvi Mikkola, Pia Trrm~inen, Aira Ollonen and Annikki Varjo for their technical assistance, and Sirkka-Liisa Leinonen for revising the English text. The work was supported by the Finnish Forest and Parks Service and the Finnish Forest Research Institute.

References B~ick J, Huttunen S, Kristen U (1993) Carbohydrate distribution and cellular injuries in acid rain and cold-treated spruce needles. Trees 8:75-84 Bergsten U (1987) Incubation of Pinus sylvestris L. and Picea abies L. (Karst.) seeds at controlled moisture content as an invigoration step in the IDS method. Dissertation. Swedish University of Agricultural Sciences. Department of Silviculture. Umeh Coulter JM, Chamberlain CJ (1910) Morphology of gymnosperms. The University of Chicago Press, Chicago Downie B, Wang BSP (1992) Upgrading germinability and vigour of jack pine, lodgepole pine, and white spruce by the IDS technique. Can J For Res 22:1124-1131 Fahn A (1990) Plant anatomy. 4th edn. Pergamon Press, Oxford Ferguson M (1904) Contributions to the life history of Pinus, with special reference to sporogenesis, the development of the gametophytes, and fertilization. Proc Wash Acad Sci 6:1-202

Fink S (1991) Unusual pattern in the distribution of calcium oxalate in spruce needles and their possible relationship to the impact of pollutants. New Phytol 119:41-51 Goo M (1951) Water absorption by tree seeds. Bull Tokyo Univ For 39:55-60 Grzywacz AP, Rosochacka J (1980) The colour of Pinus sylvestris L. seed and their susceptibility to damping-off. II. Colour of seed coats and their chemical composition. Eur J For Pathol 10:193-201 Hagner M (1985) Germinant morphology and seedling establishment in Picea abies. (in Swedish) Sver Skogsvhrdsfrr Tidsk 1:77-81 H~kansson A (1956) Seed development of Picea abies and Pinus silvestris. Medd Statens Skogsforskningsinst 46:1-23 Hoff RJ (1987) Dormancy in Pinus monticola seed related to stratification time, seed coat and genetics. Can J For Res 17:294-298 Humphrey CD, Pittman FE (1974) A simple methylene blue-azure II-basic fuchsin stain for epoxy-embedded tissue section. Stain Technol 49: 9-14 ISTA (1985) International seed testing rules for seed testing 1985. Seed Sci Technol 18:337-343 Jensen WA (1962) Botanical histochemistry. Principles and practice. Freeman, San Francisco Kamra SK (1967) Comparative studies on the germination of Scots pine and Norway spruce seed under different temperatures and photoperiods. Stud For Suec 51: 1-16 Kujala V (1927) Untersuchungen tiber den Bau und Keimfahigkeit von Kiefern- und Fichtensamen in Finnland. Comm Inst Quaest For Finlandiae 12:1-68 Leinonen K, Nygren M, Rita H (1993) Temperature control of germination in the seeds of Picea abies L. Karst. Scand J For Res 8:107-117 von Ltirzer E (1956) Megasporenmembranen bei einigen Cupressaceen. Grana Palynol 1:70-78 Nobbe F (1876) Handbuch der Samenkunde. Wiegandt, Hempel and Paren, Berlin Sarvas R (1962) Investigations on the flowering and seed crop of Pinus sylvestris. Commun Inst For Fenn 53:1 - 198 Sarvas R (1964) Conifers [Harupunt] (in Finnish) WSOY, Porvoo Schnarf K (1933) Embryologie der Gymnospennen. Handbuch der Pflanzenanatomie. X/2. Borntraeger, Berlin Schnarf K (1937) Anatomie der Gymnospermen-Samen.Handbuch der Pflanzenanatomie. X/I. Borntraeger, Berlin Singh H (1978) Embryology of gymnosperms. Encyclopedia of plant anatomy, X/2. Borntraeger, Berlin Thomson RB (1905) The megaspore membrane of the gymnosperms. Univ Toronto Stud Biol Ser 4: 85-146 Tillman-Sutela E, Kauppi A (1995) The morphological background to imbibition in seeds of Pinus sylvestris L. of different provenances. Trees 9:123-133 Voroshilova GI (1983) Formation of the wing in conifer seeds. (in Russian) Biol Nauki (Moscow) 8:81-84 Zentsch W (1960) Zur Wasseraufnahme keimender Koniferensamen. Naturwissenschaften 47:70 Zentsch W (1962) Zur Wasseraufnahme keimender Fichtensamen [Picea abies (L.) Karst.]. Flora 152:227-235