Giant Shoot Apical Meristems in Cacti Have Ordinary Leaf Primordia but Altered Phyllotaxy and Shoot Diameter

Abstract

• Background and Aims Shoot apical meristems (SAMs) in most seed plants are quite uniform in size and zonation, and molecular genetic studies of Arabidopsis and other model plants are revealing details of SAM morphogenesis. Some cacti have SAMs much larger than those of A. thaliana and other seed plants. This study examined how SAM size affects leaf primordium (LP) size, phyllotaxy and shoot diameter.

• Methods. Apices from 183 species of cacti were fixed, microtomed and studied by light microscopy.

• Key Results Cactus SAM diameter varies from 93 to 2565 µm, the latter being 36 times wider than SAMs of A. thaliana and having a volume 45 thousand times larger. Phyllotaxy ranges from distichous to having 56 rows of leaves and is not restricted to Fibonacci numbers. Leaf primordium diameter ranges from 44 to 402 µm, each encompassing many more cells than do LP of other plants. Species with high phyllotaxy have smaller LP, although the correlation is weak. There is almost no correlation between SAM diameter and LP size, but SAM diameter is strongly correlated with shoot diameter, with shoots being about 189·5 times wider than SAMs.

• Conclusions Presumably, genes such as SHOOT‐MERISTEMLESS, WUSCHEL and CLAVATA must control much larger volumes of SAM tissue in cacti than they do in A. thaliana, and genes such as PERIANTHIA might establish much more extensive fields of inhibition around LP. These giant SAMs should make it possible to more accurately map gene expression patterns relative to SAM zonation and LP sites.

INTRODUCTION

The shoot apex is the site in which a shoot’s morphology is established. Leaf primordium initiation at the base of the shoot apical meristem (SAM) establishes phyllotaxy, while spatial relationships of pith, vascular bundles, cortex and epidermis are laid out in the subapical meristem. The size of the SAM results from a balance between production of new stem cells within the meristem and loss of cells to production of leaf primordia, nodes and internodes (Howell, 1998; Lyndon, 1998). The roles of several genes are becoming understood (Nakajima and Benfey, 2002). SHOOT‐MERISTEMLESS (STM) is expressed throughout the SAM except at sites of incipient leaf primordia at its base; STM appears either to inhibit cells from differentiating or to promote cell proliferation within the centre of the SAM (Endrizzi et al., 1996). A negative feed‐back loop between WUSCHEL (WUS) and three CLAVATA (CLV) genes prevents fluctuations in SAM size during development (Weigel and Jürgens, 2002). WUS encodes a homeodomain transcription factor and is expressed in a small group of central cells two or three layers below the tunica; it confers stem cell fate on surrounding cells (Mayer et al., 1998; Schoof et al., 2000). CLV3 is expressed in the central zone (CZ) and CLV1 is expressed in the centre of the rib zone (RZ). CLV products form a complex that down‐regulates WUS. Thus if the SAM enlarges, the increased number of cells causes an increase in CLV peptides and repression of WUS, which in turn reduces the number of stem cells and SAM size. On the other hand, if too many SAM cells are incorporated into leaf primordia, the diminished number of SAM cells results in less CLV peptide, less repression of WUS, and thus a restoration of SAM size. In floral meristems (FMs), which produce a determinate number of lateral organs, conversion of SAM cells to primordium cells causes a decrease in SAM size that does not trigger a compensatory production of stem cells. Apparently AGAMOUS is active in FMs and suppresses WUS despite the reduction in levels of CLV peptide (Lenhard et al., 2001).

Much less is known about the molecular genetics and physiology controlling the location and size of leaf primordia (Laufs et al., 1998; Chuang et al., 1999). In wild‐type plants, leaf primordia never form on the summit of the SAM itself, perhaps due to STM expression. Existing leaf primordia are believed to surround themselves with overlapping fields of inhibition which determine the sites available to new leaf primordia. Wherever cells find themselves far enough from the SAM summit and in a region not inhibited by neighbouring leaf primordia, they initiate a leaf primordium (Howell, 1998; Lyndon 1998). Auxin appears to be involved (Reinhardt et al., 2000, 2003) and the gene PERIANTHIA (PAN) appears to affect the spatial extent of inhibition (Chuang et al., 1999).

The ratio of periclinal to anticlinal divisions influences the diameter of the shoot produced by the SAM. In angiosperms, the tunica consists of one or more cell layers in which periclinal divisions virtually never occur, consequently each tunica layer contributes only a single layer to the shoot primary body. Both periclinal and anticlinal divisions occur throughout the corpus but anticlinal divisions are much more common, resulting in long, slender shoots (Lyndon, 1998).

These SAM morphogenesis mechanisms appear to be flexible during development but resistant to evolutionary modifications. SAMs are typically small in seedlings but enlarge as the first few leaves are initiated (Mauseth, 1978a, 1979; Medford et al., 1992), then they often increase greatly if they switch to acting as inflorescence meristems. As SAMs enlarge, phyllotaxy and shoot width usually increase, sometimes spectacularly as in the transition from a slender sunflower shoot to a broad capitulum, both produced by the same SAM (Hernandez and Palmer, 1988). A similarly dramatic, long‐term transition occurs during the juvenile–adult phase change of Melocactus: a single SAM produces an ordinary juvenile cactus body for years then switches to producing a flower‐bearing cephalium for the rest of the plant’s life (Niklas and Mauseth, 1981; Mauseth, 1989). Conversely, there seems to be little evolutionary variation in SAMs. The range of sizes for vegetative SAMs is restricted in seed plants, with almost all conifers and angiosperms having SAM diameters between 100 and 300 µm (Johnson, 1951; Gifford, 1954; Mauseth, 1988). Similarly, phyllotaxy is low in most taxa and shoots are narrow.

However, extreme modifications of the shoot apex occurred during the evolution of the cactus family. Shoot diameters range from slender to exceptionally wide (up to 80 cm diameter in Echinocactus grusonii and E. platyacanthus; Anderson, 2001; Mauseth, 2000; Mauseth et al., 2002), SAMs vary in diameter from 80 to 1500 µm, some of the largest known (Boke, 1957; Mauseth, 1978b), and phyllotaxy varies from distichous to having as many as 120 orthostichies (ribs) (Robberecht and Nobel, 1983; Gibson and Nobel, 1986; Terrazas and Mauseth, 2002). Although often described as leafless, all cacti produce leaves which, despite being very small in most species, are relatively ordinary structurally in that most have vascular tissue, stomata and dorsiventral organization (J. D. Mauseth, unpubl. res.). The exceptional ranges of shoot apex features are almost certainly due to altered expression patterns and interactions between STM, WUS, CLV and other genes. The giant cactus SAMs may provide valuable research material for examining the molecular genetics of SAM morphogenesis. This paper presents the results of initial studies as to how the evolution of giant SAMs has affected the development of leaf primordia, phyllotaxy and shoot diameter.

MATERIALS AND METHODS

Samples of 183 species were examined from mature plants; seedling SAMs were not studied. Combined with species studied previously, the total sample was 205 species. Two or three SAMs were sampled for most species, four samples were obtained for some. Subfamily Opuntioideae was not included because their shoots have determinate growth: their SAMs function only briefly then disorganize. Most samples were collected from adult plants in habitat, others were obtained from commercial nurseries (Abbey Garden, La Habra, CA, USA; Mesa Garden, Belen, NM, USA).

Phyllotactic rows are readily identifiable on many cactus shoots (Mauseth, 2000). Whereas leaves remain microscopic even when fully developed, each spine cluster is a modified axillary bud, thus indicating the site where a leaf primordium was formed (Fig. 1; Mauseth et al., 2002; Terrazas and Mauseth, 2002). Phyllotactic rows are easiest to identify in columnar cacti because vertically aligned leaf primordia merge together forming ribs (orthostichies) that run up the shoot. Species that have a globose form may also have ribs but they more often have parastichies visible as intersecting spirals of protuberances called tubercles. Each tubercle is a bit of shoot that expands to be a conical projection just below a leaf primordium.

Fig. 1. Cactus SAMs. (Note that the same magnification is used for Figs 1A, D–F and 2A–C.) (A and B) Oroya peruviana. (A) Longitudinal section of shoot apex and an orthostichy of leaf primordia that are forming a rib. Arrows indicate four leaves (rightmost is just a primordium); S, a young spine being produced by an axillary bud; cbs, developing system of cortical bundles. Material above the SAM and leaf primordia consists of trichomes. Long bar = 500 µm, short bar = 70 µm, the width of the SAM of Arabidopsis thaliana for comparison. (B) Magnification of part (A), showing that all SAM zones are large with large numbers of cells. Scale bar = 100 µm. (C) High magnification of part (A), showing the transition area between the CZ and PZ. Arrows indicate two periclinal divisions in the second layer of cells, indicating that these cells are part of the CZ; the tunica is only one cell layer thick. Scale bar = 100 µm. (D) Rhipsalis heteroclada, with one of the smallest SAMs found in cacti. Scale bar = 500 µm. (E) Pereskia sacharosa, a leafy, nonsucculent cactus with many relictual features and a small SAM. Scale bar = 500 µm. (F) Echinocactus platyacanthus, with the largest known SAM. Arrows indicate two leaf primordia. Long bar = 500 µm, short bar above the SAM in both (F) and (G) is 70 µm, the width of the SAM of A. thaliana. (G) E. platyacanthus, magnification of CZ–PZ transition area, showing the very large numbers of cells present in each zone. Scale bar = 500 µm.

Apices were fixed in Navashin’s solution, dehydrated through a tertiary‐butanol series, then embedded in Paraplast Plus. Sections were stained with Safranin and Fast Green (Mauseth et al., 1984). SAM diameter at the level of the youngest leaf primordia was measured on median longitudinal sections using a light microscope, then the circumference of the SAM base was calculated. Leaf primordium size was calculated by dividing SAM base circumference by the number of orthostichies in columnar species or by the larger of the two parastichies in globose species. Shoot diameter was measured from the base of the ribs or tubercles on one side of the shoot to the base of those on the other side. Ribs and tubercles were excluded because they are not produced directly by the peripheral zone (PZ) of the SAM as the rest of the cortex is; instead they undergo an additional period of cell proliferation and may even have their own meristem (J. D. Mauseth, unpubl. res.).

RESULTS

Zonation

All SAMs in all species had the same zonation, despite great differences in meristem size (Figs 1A, B and F and 2A–D). The tunica was always unistratose; in many samples it appeared multistratose on the periphery where it overlay the PZ, but if the presumed inner layers were followed carefully across the very summit of the SAM, they were irregular at some point, indicating that they were actually part of the PZ and CZ rather than being true tunica (Fig. 1C). The corpus always had three zones, CZ, PZ and RZ. Every zone was larger in larger SAMs (Figs 1G and 2D); there was no case of a meristem being larger due to just a single zone being larger while the other three remained small.

Fig. 2. Giant meristems: (A) Gymnocalycium saglionis; (B) Lobivia ferox; (C) Matucana aurantiaca; (D) Trichocereus candicans. Scale bar = 500 µm in all micrographs.

SAM size

Six of the newly examined species had SAMs wider than 1500 µm, the upper limit of the range established by earlier studies. Many species had diameters at the high end of the range (Table 1). The narrowest cactus SAM had a diameter of 93 µm in Rhipsalis crispata, the widest was 2565 µm in Echinocactus platyacanthus. Soehrensia korethroides was also extremely large, 2032 µm diameter (Fig. 1D–G). Combining data for all 205 species, mean diameter was 639 µm.

Table 1.

Newly studied species with SAM diameters greater than 900 µm

Species	SAM diameter (µm)	No. of orthostichies and parastichies
Trichocereus candicans	900	10
Echinofossulocactus species (Lau 738)†	924	54
Echinopsis leucantha	924	17
Haageocereus acranthus	924	12
Echinocereus pectinatus	959	13
Azureocereus hertlingianus	960	17
Escontria chiotilla	960	7
Espostoa lanata	960	27
Espostoa lanata	960	18
Matucana grandiflora	960	14
Oroya depressa	960	28
Parodia stuemeri	960	33
Stetsonia coryne	960	10
Thelocactus lophothele	960	13
Pyrrhocactus villicumensis	965	10
Matucana formosa	972	23
Solisia pectinifera	1000	13
Browningia candelaris	1020	21
Matucana aurantiaca	1020	27
Parodia maassii	1080	17
Soehrensia formosa	1116	20
Buiningia aurea	1140	14
Pyrrhocactus bulbocalyx	1143	18
Melocactus peruvianus, cephalium	1200	*
Thelocactus rinconensis	1200	21
Trichocereus pasacana	1320	18
Ferocactus chrysacanthus	1387	15
Gymnocalycium asterias	1404	15
Soehrensia bruchii	1425	32
Lobivia ferox	1440	30
Sulcorebutia mentosa	1440	*
Mammillaria heyderi	1500	*
Echinocactus platyacanthus	1680	21
Brachycalycium tilcarense	1710	29
Echinocactus grusonii	1852	33
Echinocactus platyacanthus	1860	21
Soehrensia korethroides	2032	30
Echinocactus platyacanthus	2565	36

* Phyllotactic spirals could not be identified with sufficient certainty to permit counting.

^† This is an unnamed species known scientifically and commercially by this name and number.

SAM diameter and phyllotaxy

The number of ribs (vertical orthostichies) on a shoot was not restricted to being a Fibonacci number (1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144). Every possible number of ribs from two to 36 was found in the sampled cacti, and two species of Echinofossulocactus had 54 ribs, one had 56 but none had the Fibonacci number of 55 (Tables 1 and 2). In many cacti with globose bodies, left‐ and right‐handed spiral parastichies were readily apparent and these frequently corresponded to Fibonacci numbers. Phyllotaxies of 5+8, 8+13, and 13+21 were most common. Phyllotaxy was often irregular because new parastichies were added as the body continued to grow into an ever larger sphere.

Table 2.

Species with exceptionally small leaf primordia (<100 µm diameter)

Species	SAM diameter (µm)	No. of orthostichies and parastichies	Leaf primordium diameter (µm)
Echinofossulocactus multicostatus	780	56	44
Echinofossulocactus species     (Lau 738)	840	56	47
Aporocactus flagelliformis	208	10	65
Cleistocactus strausii	624	30	65
Chamaecereus silvestris	181	8	71
Selenicereus donkelarii	208	9	73
Haageocereus versicolor	480	19	79
Monvillea kronleinii	229	9	80
Micranthocereus polyanthus	360	14	81
Cleistocactus strausii	780	30	82
Bolivicereus serpens	262	10	82
Uebelmannia meninensis	840	32	82
Trichocereus strigosus	482	18	84
Rebutia wessneriana	686	24	90
Parodia stuemeri	960	33	91
Escobaria sneedii	382	13	92
Haageocereus pseudoversicolor	600	20	94
Selenicereus testudo	210	7	94
Matucana hystrix	780	26	94
Uebelmannia gummifera	825	27	96
Matucana cereoides	876	28	98

SAM diameter was partially correlated with the number of orthostichies or parastichies (Fig. 3). Plants with small SAMs between 100 and 400 µm diameter had only about two to 16 orthostichies or parastichies, whereas those with large SAMs of 1500 µm or wider had between 21 and 36. For species with SAMs of intermediate size, the number of orthostichies and parastichies was extremely variable, ranging from as few as five to as many as 56.

Fig. 3. SAM diameter in µm versus the number of orthostichies or parastichies. The four points labelled ‘1’ are all Echinofossulocactus, which have unusually high numbers of orthostichies (36–54) produced by relatively small SAMs. The two points labelled ‘2’ are Neoraimondia, which have low numbers of orthostichies (five or six) despite having SAMs about the size of those of Echinofossulocactus.

The possibility of experimental error causing the large variability in Fig. 3 was examined by rechecking all exceptional values and by obtaining additional samples of outlying species.

Experimental error could account for only a small portion of the variability. All four apices at point 1 are Echinofossulocactus; for the position of these apices to be due to experimental error, the plants would need to have only 14 ribs instead of 34–56 or their SAMs would have to be about 3400 µm in diameter instead of 600 µm. For point 2, the plants (both Neoraimondia) would need to have either 13 ribs or a SAM diameter of <0·0 µm; but having only five or six ribs is diagnostic of these species, and their SAM diameters are indeed not negative numbers. Counting ribs and phyllotactic spirals was easy and accurate, so that should have introduced little error. Mismeasuring SAM width by even large amounts, for example 10 % or 50 µm in small apices or 100–250 µm in larger ones, would shift most points only slightly, so any mismeasurement contributed only slightly to the variability in Fig. 3. Instead, leaf primordium size differed among species.

SAM diameter and leaf primordium diameter

Leaf primordium size varied markedly and was only slightly correlated with SAM diameter (Fig. 4 and Tables 2 and 3). Species with SAMs of the same size had significantly different numbers of rows of leaf primordia being initiated at their base, which required that leaf primordia differ significantly in size. SAMs with diameters between 500 and 1000 µm could have leaf primordia as small as about 50 µm across or almost ten times as large, 400 µm. Conversely, many of the narrowest SAMs on the left side of Fig. 4 had leaf primordia larger than those of the most gigantic SAM on the right side. Mean leaf primordium diameter was 169 µm, minimum was 44 µm and maximum was 402 µm.

Fig. 4. SAM diameter versus leaf primordium diameter, both in µm.

Table 3.

Species with exceptionally large leaf primordia (greater than 250 µm diameter)

Species	SAM diameter (µm)	No. of orthostichies and parastichies	Leaf primordium diameter (µm)
Echinocactus platyacanthus	1680	21	251
Buiningia aurea	1140	14	256
Castellanosia caineana	660	8	259
Neoraimondia roseiflora	420	5	264
Thelocactus heterochroma	672	8	264
Pachycereus weberi	840	10	264
Echinocereus enneacanthus	621	7	279
Trichocereus puquiensis	712	8	280
Browningia pilleifera	720	8	283
Trichocereus candicans	900	10	283
Ferocactus chrysacanthus	1387	15	290
Gymnocalycium asterium	1404	15	294
Escontria chiotilla	660	7	296
Stetsonia coryne	960	10	302
Neoraimondia roseiflora	624	5	392
Neoraimondia herzogiana	768	6	402

Phyllotaxy and leaf primordium diameter

There was a negative correlation between phyllotaxy and leaf primordium diameter: species with greater numbers of orthostichies or parastichies had smaller leaf primordia (Fig. 5). Variability was great for species with 18 or fewer ribs or parastichies. For example, species with about ten ribs had leaf primordia varying from about 60 µm in diameter to over 300 µm (Tables 2 and 3).

Fig. 5. Leaf primordium diameter in µm versus number of orthostichies or parastichies.

SAM diameter and shoot diameter

Shoot diameter was strongly correlated with SAM diameter (Fig. 6). There was very little variation among the samples except that four (Oreocereus celsianus, Soehrensia formosa, Echinocactus grusonii and E. platyacanthus, points 1–4, respectively) had exceptionally wide shoots relative to their SAMs, and one shoot (Buiningia aurea, point 5) was somewhat narrower than expected. The extremely broad shoots were noteworthy for having such exceptional diameters: shoot diameter of 662 mm in E. platyacanthus, 435 mm in E. grusonii, and 369 mm in S. formosa. For comparison, shoots (the inflorescence stalks) of A. thaliana were 0·5 mm thick with a cross‐sectional area of 0·2 mm² (J. D. Mauseth, unpubl. res.), so shoots of E. platyacanthus were 1324 times wider and had a cross‐sectional area of 344 196 mm² or 1 720 980 times larger. The large diameter of cactus shoots was almost completely due to the primary tissues of pith and cortex because in all specimens wood was at most 7 mm thick. However, many cacti had narrow shoots with diameters less than 20·0 mm; these invariably had narrow SAMs (Fig. 1D).

Fig. 6. SAM diameter in µm versus shoot diameter in mm.

The regression equation indicated that the mature shoot was approx. 189·5 times wider than the base of the SAM.

DISCUSSION

Apices of A. thaliana are being studied intensively, and almost certainly the SAM control mechanisms being found there also occur in most or all other seed plants, including cacti. If so, there must have been significant modifications of some of these genes during the evolution of the giant meristems of cacti. For example, the A. thaliana SAM is 70 µm in diameter (Medford et al., 1992), whereas that Echinocactus platyacanthus is 2565 µm; if both are approximately hemispherical, then the volume of an A. thaliana SAM is 9 × 10⁴ µm³ and that of E. platyacanthus is 4 × 10⁹ µm³, about 45 thousand times larger (Fig. 1A and F, small scale bars). This tremendous difference raises questions about the spatial expression patterns of genes. If STM is responsible for inhibiting cells from differentiating, it would have to control a vastly greater volume in the giant SAMs of cacti as opposed to the smaller A. thaliana apices. But control could be effected by several mechanisms. Perhaps all SAM cells express STM and produce STM peptide at about the same rate as in A. thaliana cells. Alternatively STM may be expressed by about the same number of cells as in A. thaliana but with each cell expressing at a high level, producing enough STM that it can be transported throughout the large volume of a giant SAM. It is also possible that STM needs to inhibit differentiation only during the four or five rounds of cell division immediately preceding leaf primordium initiation; this would encompass more or less all cells of the small A. thaliana SAM but would be just a plate of cells about 70 µm thick at the base of a giant SAM.

Similarly, there are several ways that the WUS–CLV feed‐back loop could have been modified during evolution. In A. thaliana, WUS mRNA is expressed in a small group of cells in the CZ (Mayer et al., 1998; Schoof et al., 2000) but not throughout the entire CZ. The giant SAMs in cacti have enlarged CZs (Figs 1B and F and 2D; relative volumes of corpus zones are constant in cacti; Mauseth and Niklas, 1979), but because the CZ histological boundary does not coincide with the WUS‐expression boundary, it is not known if the WUS‐expressing group of cells has been enlarged during evolution. Giant SAMs could have a greater number of WUS‐expressing cells or they could have just a small number that is especially active. Alternatively, it could be that the number and activity of WUS‐expressing cells have remained constant during cactus evolution but the surrounding cells have become more sensitive to their effect. Evidence favouring the former hypothesis is provided by fasciated SAMs produced by mutating CLV genes: the resulting SAMs were plate‐like, being extremely wide in one direction but of ordinary height and width (Schoof et al., 2000). In these clv‐fasciated SAMs, WUS‐expression occurred in a very wide but still shallow band that extended the full width of the SAM: its dimensions remained proportional to those of the SAM. Consequently, it may be that in cactus giant SAMs, WUS expression will occur as a greatly enlarged sphere of cells in the CZ with dimensions proportional to those of the SAM.

CLV expression is more difficult to predict. In A. thaliana, CLV3 is expressed in cells above and to the sides of the WUS‐expression domain, and CLV1 is expressed in the centre of the RZ. The two expression domains overlap only slightly even though their gene products must interact to control WUS (Clark et al., 1997; Fletcher et al., 1999). If this same expression pattern occurs in giant SAMs, then the CLV1‐expression domain in the RZ might not overlap at all with that of CLV3 located at the top of the CZ: in a large SAM the two domains might be separated by hundreds of micrometres. However, the location of the CLV1‐expression domain in the RZ of A. thaliana may just be coincidental rather than a specific function of the RZ: perhaps the important criterion is that CLV1 expression be immediately proximal to the WUS‐expression domain, which due to the small size of A. thaliana SAMs, happens to be in the RZ. Perhaps in giant meristems it too would be located in the CZ. Studies of these genes in cactus giant SAMs would be valuable not only for examining the universality of these apex control mechanisms but also for clarifying their spatial relationships.

Within some cacti, leaf primordia evolved to be very large, others have a more ordinary size: the relative sizes of leaf primordia and SAMs are not closely correlated. However, most species with really large SAMs also have exceptionally large leaf primordia (Fig. 4). It may be important to distinguish between the set of leaf primordium founder cells and the inhibition zone surrounding each primordium. In Nicotiana tabacum there is a set of about 150 founder cells, distributed in the outermost three layers of the PZ and forming a plate about 12–13 cells wide (Poethig and Sussex 1985a, b). Impatiens has about 100 cells (Battey and Lyndon, 1988) and Zea about 200 (Poethig, 1984). Although most cactus leaves remain microscopically small when fully formed, most develop chlorophyllous mesophyll and a vascular bundle (Mauseth, 1977; J. D. Mauseth, unpubl. res.). Consequently, cactus leaf primordia are probably just as complex as those of other dicots. But there is a ten‐fold range in leaf primordium size in cacti (from 44 to 402 µm) that must be explained. It might be that large leaf primordia have a set of founder cells 100 times larger than small leaf primordia. Or the number of founder cells may be more or less the same but ‘large primordia’ are really just primordia with an ordinary number of founder cells surrounded by a large field of inhibition. The latter hypothesis is supported by evidence that PERIANTHIA (PAN) affects primordium spacing by facilitating the diffusion, reception or production of a leaf‐initiation inhibitor (Chuang et al., 1999). Thus what appear to be large leaf primordia in cacti might be an ordinary number of founder cells with an especially active expression of PAN.

The strong correlation between SAM diameter and shoot diameter indicates that production of the exceptionally broad, succulent cactus shoots may be a direct result of having large SAMs, without any evolutionary alteration of the ratio between periclinal and anticlinal divisions. Because the relative volumes of zones are constant in cacti, giant apices have giant zones (Mauseth and Niklas, 1979). If cell size remains constant, then as SAM size increases the number of cell layers in the PZ increases. Even if the ratio of periclinal to anticlinal divisions remains the same within each layer of PZ cells, doubling the number of layers doubles the number of periclinal divisions; a cortex produced by a thick PZ will be wider relative to its length as opposed to a cortex produced by a thin PZ. Increased SAM size leads to increased numbers of cell layers in another way as well. Although cell division may be slower in the CZ (Mauseth, 1976; Lyndon, 1998), division must occur throughout the PZ not merely to the point at which leaf primordia are initiated but even below that, down to a level at which all the cells of nodes and internodes have been produced. Therefore, large SAMs have not only thicker PZs but also longer PZs and thus more opportunity for periclinal divisions to add cell layers to the cortex. The same relationships apply to the RZ and pith as well.

Much remains to be learned about SAM genes such as STM, WUS, CLV, PAN and others, and cacti may provide especially useful research material. By studying the activity of these genes in the giant SAMs of cacti, it should be possible to more accurately map their expression patterns in relationship to SAM zones and to sites of leaf primordium initiation. Also, because cactus seedlings have only small apices (Mauseth, 1978a, 1979), their SAMs must enlarge greatly during vegetative development, making it possible to examine gene expression and activity during meristem enlargement and in meristems of greatly different sizes within the same species. Many of the cacti in Tables 1–3 are readily available from commercial nurseries and should be useful study material.

ACKNOWLEDGEMENTS

I thank the Lozano Long Institute of Latin American Studies at the University of Texas for Mellon Fellowships to support this research, and the US National Park Service for permission to collect in Big Bend National Park.

Received: 21 January 2004; Returned for revision: 18 February 2004; Accepted: 12 March 2004. Published electronically: 14 May 2004