Laticifers and secretory ducts: two other tube systems in plants


  • 1

    A laticifer is thus named because it contains a latex. ‘Latex’ derives from poetical Latin and denotes a (frequently special) type of liquid; the Latin plural of ‘latex’ is ‘latices’. In its general botanical sense, latex is of highly variable chemical composition, is not necessarily of a milky appearance, and may contain precipitates of various sorts, or suspended colloids, or a variety of dissolved solutes (cf. Esau, 1965; Mahlberg, 1993). Moreover, because latex frequently is milky in appearance, a tissue that contains it may occasionally be termed ‘lactiferous’ (e.g. Haug et al., 2005); and here the ‘lact’ comes from a Latin root denoting breast milk. This author inclines toward the view that, if a tissue is known to contain laticifers in the sense of Esau (1965), it is certainly laticiferous, whereas, if a tissue is known only to exude upon puncture a whitish liquid, then it is still laticiferous because its exudate (however stored in the tissue) is certainly not breast milk.

  • 2

    Symplastic transport itself is more a distributed mechanism than one concentrated in a supracellular anatomical feature (i.e. tube), and it was therefore placed beyond the scope of this survey. However, recent relevant reviews have treated the significance of plasmodesmograms (Van Bel & Oparka, 1995); the role of plasmodesmata in symplastic transport (Roberts & Oparka, 2003); long-distance transport in nonvascular plants (Raven, 2003); the role of cytoplasmic streaming in symplastic transport (Pickard, 2003); and macromolecular trafficking (Oparka, 2004).

  • 3

    With the flowering of molecular taxonomy, the structure of taxonomic trees underwent and is still undergoing a period of great change. In this paper the tendency is to adhere to traditional nomenclature but within a framework of the tree structure continuously updated by the Angiosperm Phylogeny Website ( hosted by the Missouri Botanical Garden. The URLs and are also of value.

  • 4

    Exactly how to differentiate these plant products from one another is something of a judgement call (cf. Nair, 1995). However, polysaccharides are generally the prominent component in a gum. A gum will dissolve in (or form a suspension with) water to yield a viscous (and often adhesive) jelly, paste, or syrup. Gums are insoluble in organic solvents of low dielectric constant. Resins, in contrast, are classified as aromatic because of the prominence of ring compounds (often heavy and nonvolatile) in them. Resins are not water soluble but dissolve readily in organic solvents of low dielectric constant. Essential oils are lower molecular weight aromatic compounds of a distinctive odor (essence) which characterizes their source; they are a common constituent of resins.

  • 5

    In modern usage, the ‘apoplast constitutes all compartments beyond the plasmalemma’ (Sattelmacher, 2001, p. 167) while, by exclusion, the symplast seems to include everything within the plasmalemma. This view, while formally correct, does not encompass the normally unvoiced expectation that both these spaces should be filled with electrolyte of unremarkable properties; and a laticifer vacuole filled with rubber particles or a duct loaded with terpenoids may require rethinking some generalizations.

Author for correspondence: William F. Pickard Tel: +1 314 935 6104Fax: +1 314 935 7500Email:


The plant kingdom has elaborated several conducting systems. Three are primarily for mass transport: the aerenchyma (for gas exchange in submerged parts), the phloem (for exchange of nutrients within the plant), and the xylem (largely for transport of water from soil to transpiring leaves). Two others are believed to be primarily defensive and to store under pressure aversive contents which they exude when punctured: the laticifer and the secretory duct. This review provides for the latter two systems the highlights of what is known about their general physiology and ecophysiology but not their metabolism and their molecular biology. It is argued that, given the importance of laticifers and secretory ducts to plant defense against insect herbivory, these structures are under-investigated and deserve more intensive study.


Vascular plants (tracheophytes) are called ‘vascular’ because they are characterized by two anatomical ‘tubing’ systems whose presence is virtually universal: the prominent tracheal-appearing xylem, which transports soil water to regions within the plant where its chemical potential is lower, especially the upper plant; and the less obvious phloem, which transports photosynthate from leaves to heterotrophic tissues. Frequently these two are supplemented by a third, less common, system: the aerenchyma, which is elongate but not tubular and facilitates internal gas transport. All three of these transport systems, without which plants would be unable to perform the primary metabolic activities necessary for survival, have been widely studied, both classically and in recent times (defined to be from 1995 onwards); and the study of all three systems has profited greatly from the explosion of experimental techniques which has characterized the past few decades.

Xylem is a popular topic of research and, in recent years, has been a subject of over 4900 papers, including some 190 reviews. These reviews covered such diverse topics as hydraulics and turgor (e.g. Sack & Holbrook, 2006), signaling (e.g. Starck, 2006), solute movement (e.g. Canny, 1995; Sattelmacher, 2001; De Boer & Volkov, 2003), and transport mechanisms (e.g. Steudle, 2001). Water movement in the xylem has been heavily studied and reviewed; and, despite some controversy in recent years, the mechanism most commonly cited to explain it is the cohesion–tension hypothesis of Dixon and Joly (e.g. Pickard & Melcher, 2005). Under this mechanism the xylem sap flows to those regions of a xylem domain where the free energy of the sap water is lowest, normally the apoplast directly under a stoma; even under idealized circumstances the mathematical description of this movement is involved (cf. Pickard, 1981).

Phloem, like xylem, is a popular topic of research and, in recent years, has been a subject of over 3000 papers, including some 160 reviews. These reviews covered such topics as the strategies and mechanisms of phloem unloading (e.g. Oparka & Cruz, 2000), signaling and macromolecular trafficking (e.g. Starck, 2006), modeling (e.g. Minchin & Lacointe, 2005), general function (e.g. Van Bel et al., 2002; Van Bel, 2003); and even so specialized a topic as phloem response to aphid saliva (Miles, 1999). Phloem translocation has been heavily studied and reviewed, and is most commonly explained in terms of the pressure-driven mass-flow hypothesis of Münch (e.g. Van Bel & Hafke, 2005). Under this mechanism the phloem sap is impelled to those regions of a phloem domain where the free energy of the sap is lowest, normally roots and other heterotrophic tissue; even under idealized circumstances the mathematical description of this movement is involved (cf. Thompson, 2006).

Aerenchyma has recently been discussed from a variety of viewpoints; and it has been a factor in over 420 papers, including some 36 reviews. For example, Raven (1996) emphasized the development, function, and maintenance of gas spaces; Grosse et al. (1996) reviewed 150 yr of pressurized gas-flow studies in plants; Colmer (2003) stressed internal aeration and the mechanisms of long-distance gas transport; and Evans (2004) focussed upon aerenchyma formation by the two processes of schizogeny and lysigeny. Of general interest, for both aerenchyma and xylem, is Kozela & Regan's (2003) examination of the ability of a plant to elaborate files of cells which it then converts into tube systems through cellular autolysis. A majority of workers would agree that, in aerenchymal gas transport, the mechanism of diffusion is often supplemented by the pressure drop associated with a constriction-induced velocity increase (i.e. Venturi suction; cf. Colmer 2003, especially fig. 1) and/or molecular effusion (cf. Jeans, 1940, sections 39–41; Colmer, 2003). But, despite the activity to which these reviews give witness, there are still many incompletely answered questions, especially concerning the relative importance of these three mechanisms of gas transport within a given aerenchyma.

By contrast, anatomical structures involved in storing, moving, and physically releasing secondary metabolites are not so generally present in plants, although they are widely scattered within the Plantae, can be very specific in content, anatomy, and physiology, and are of presumptively high value to the plant. The secondary metabolites themselves are frequently viewed as weapons in a plant–herbivore arms race; and because of their characteristic scent, stickiness, or vesicant properties they are often familiar to persons otherwise unfamiliar with secondary metabolites, much less phloem or xylem. The biochemical and genetic study of these secondary metabolites – along with their utility as sources of drugs, elastomers, flavorings, and miscellaneous terpenoids – has proceeded at a fair rate (e.g. Trapp & Croteau, 2001; Facchini & St-Pierre, 2005; Facchini et al., 2005; Kutchan, 2005). Similarly, the gross anatomical details of how the plant sequesters and stores these metabolites have been elaborated considerably (e.g. Esau, 1965; Fahn, 1979), and it is now believed that they should be broadly divided into two tubing systems: the laticifers1, in which the secondary metabolites are stored inside the living cell(s) which produces them; and the secretory ducts, in which they are stored in an extracellular space. Although the solution of each taxon to these challenges has many features in common with those of other taxa, it can at the same time display a highly specific elaboration of metabolite selection, mechanisms by which the plant is protected from its own chemical defenses, and specialized adaptations for the synthesis, collection, and delivery of the secondary metabolites.

Laticifers have not recently been a popular topic of study; and the past decade has seen the appearance of only about 10 papers a year. Aspects that have been studied include their macro-physiology in Hevea (e.g. Pakianathan et al., 1989) and their role in the production of Hevea latex (d’Auzac, 1989a,b; d’Auzac et al., 1995); their probable importance in insect defense (e.g. Dussourd, 2003); and their involvement in the transport pathways of natural products (Weid et al., 2004; Kutchan, 2005). However, details of the importance, mechanism, or phenomenology of mass transport within them remain vague.

Ducts likewise are not currently a popular target of research. Recently, perhaps 10 papers a year have discussed their physiology. Only one review of likely interest to a transport-oriented physiologist was located: Francheschi et al. (2005) treated resin ducts from the viewpoint of a defense against bark beetles. To find more comprehensive reviews it is necessary to revisit the classic contributions of Fahn (e.g. 1979, 1988).

A variety of studies of secretion from one cell to another or to extracellular spaces have provided an excellent background for visualizing how secretion into laticifers or ducts might occur. But only to a limited extent are there illuminating data on the detailed composition of the luminal liquid, or its rate of production, or possible mass transport within an undamaged lumen, or the quantitative phenomena of liquid flow after damage by a herbivore. This is true not only for typical plants of indifferent ecological or economic importance but also for Hevea brasiliensis, Papaver somniferum, and Pseudotsuga menziesii. By contrast, the extensive studies that have characterized the behavior of xylem and phloem in a wide range of plants have made use not only of modern molecular biological tools but also of a wide variety of biophysical and microchemical techniques for examining, at both tissue and cellular levels, the motion of sap and its pressure, osmotic characteristics, inclusions, microchemistry, et cetera. Even the transport-related aspects of the symplast have received greater attention2 than those of the laticifer and secretory duct.

In this survey, the concentration will be upon laticifers and secretory ducts, touching upon their development and anatomy, but focussing upon their biophysics and their ecophysiology. Relatively little will be said about their biochemistry and genetics. Emphasis will be placed upon their putative roles as plant defense systems, how these tubes expel and then replenish their contents, and the effects of their contents once expelled. A principal goal of this review will be to emphasize that much, relative to xylem and phloem, is not known; and this claim will be illustrated by suggesting a series of open questions whose answers would go some way toward redressing this disparity.



Laticifers are specialized cells (or files of cells) that contain a slurry or suspension of many small particles in a sap of unspecified composition, but normally of a refractive index different from the indices of the particles (cf. Fahn, 1979; Mahlberg, 1993). Although most often milky in color, this sap may also be yellow, orange, red, brown, or even colorless. The combination of suspension plus sap is called ‘latex’. Laticifers are suspected, and in many cases known, to have roles in herbivory and/or disease resistance (e.g. Fahn, 1979; Dussourd & Eisner, 1987; Dussourd & Denno, 1991; Dussourd, 1999; Konno et al., 2004); and they are of some global economic importance, being crucial in the production of both opium and natural rubber.

Anatomy and development

The nonarticulated laticifer (as found in milkweed (Asclepias spp.) or hemp (Cannabis spp.)) is a single elongated cell which develops by intrusive growth, may be tens of centimeters long (Mahlberg, 1993; Lev-Yadun, 2001), and may be branched or unbranched. The articulated laticifer (as found in dandelion (Taraxacum spp.) or moonflower (Ipomoea alba) or the rubber tree (Hevea brasiliensis)) is a file of somewhat elongated cells, which can also extend considerable distances, and may be either nonanastomosing or anastomosing (Esau, 1965; Evert, 2006); the moonflower laticifer is illustrated in Fig. 1. Both types of laticifer tend to be associated with vascular bundles. Both are widely distributed within the tracheophytes3, with examples being found in the ferns (Labouriau, 1952), the monocots (Esau, 1965, fig. 13.6), and the dicots, especially the rosids and the asterids. Within the Euphorbiaceae, the type of laticifer varies from genus to genus (e.g. Rudall, 1994); and within the Asclepiadoideae (Apocynaceae) both types have been reported to occur within the same species (Esau, 1965, p. 321; Mahlberg, 1993, p. 14). Interestingly, neither apparent vasculature nor putative laticifers are confined to the plants; and both can be found, for example, in the fungi ( at the University of Wyoming, WY, USA).

Figure 1.

Planed longitudinal sections of a stem of moonflower (Ipomoea alba L.) viewed in transmitted white light. (a) Several files of nonanastomosing articulated laticifers; the dark regions are intracellular concentrations of latex. (b) The arrow indicates the junction of two laticifers in a file; details of the ultrastructure of this junction are unknown. The oval surrounds two of the small vesicles which are prominent in the exudate. The author thanks Professor Margaret E. McCully (then of Carlton University, Ottawa, Ontario, Canada) for extending the hospitality of her laboratory where he made these photographs.

The development of the nonarticulated laticifer has been magisterially reviewed by Mahlberg (1993); but a similar review has not been carried out for articulated laticifers, and Mahlberg's conclusions have not been notably extended in the years since his publication appeared. Nonarticulated laticifers are coenocytic and appear to become longer through karyokinesis and intrusive growth of the cell (Serpe et al., 2002), a process that possibly occurs by schizogeny as a small bleb expands into an intercellular space and then greatly expands that space by increasing its turgor; interestingly, this hypothesized process – if in fact it exists – is not associated with callose production, even though schizogeny might have been predicted to trigger defense or wound responses (cf. Serpe et al., 2002). Articulated laticifers commonly appear in the embryo but are extended during growth by conversion of more apical meristematic cells into laticifers rather than by intrusive growth of the embryonic primordia (Esau, 1965).

Mature nonarticulated laticifers are thought to have few if any plasmodesmata (e.g. Serpe et al., 2002); neither is there evidence for plasmodesmata in mature articulated laticifers (e.g. Fisher, 1991). This, of course, is not necessarily true of early and developing laticifers (Inamdar et al., 1988). While formal plasmodesmogram-type analyses of laticiferous tissue seem never to have been performed and detailed electron microscopic analyses are few in number, it nonetheless appears that a mature laticifer often forms its own isolated symplastic domain, unconnected with its neighboring parenchyma by plasmodesmata (e.g. de Faÿet al., 1989a,b).

General physiology

The single most obvious characteristic of a laticifer that distinguishes it from other cells is its internal latex, which oozes copiously whenever the laticifer is punctured; and the unpunctured laticifer presumably is turgid as a result of osmotic water uptake. Therefore, it seems only reasonable to assume that, upon puncture of the laticifer, the efflux will display an initial pressure-driven surge which grades rapidly into a slower osmotically driven flow as the osmotic contents of the laticifer are gradually diluted by water uptake from an apoplast of much lower osmolality, and mechanisms of latex coagulation, operating on a much longer time scale, gradually plug the laticifer. The existence of gelling/plugging/clogging phenomena in phloem, in resin ducts, and in laticifers seems accepted (e.g. Dussourd & Denno, 1991, pp. 1384–1385; Dussourd, 1999, p. 511). Nevertheless, the author, although familiar with the hardening of kauri resin and with the heady odor of a freshly coagulated cup lump on a rubber tree, was unable to locate the extensive literature on mechanisms that surely would be needed to elucidate these phenomena.

The commercial realities of flow from groups of Hevea laticifers have of course been investigated extensively (cf. Pakianathan et al., 1989) as has the measurement of pressure associated with puncture of a region of laticiferous tissue (e.g. Milburn & Ranasinghe, 1996). Time constants of latex flow in favored Hevea clones can be on the order of many tens of minutes; but similar data on taxa that have not been subjected to intensive selective breeding remain largely uncollected. In H. brasiliensis, the pressure from groups of laticifers measured manometrically lies in the range 0.5–1.0 MPa, while osmolality measurements on the latex give osmotic potentials at or slightly above 1 MPa (Milburn & Ranasinghe, 1996); other studies of Hevea point to potassium (K+), sucrose, malate, and amino acids as the principal solute species, a total osmolality on the order of 200 mOsm, and an inferred osmotic pressure of c. 0.5 MPa (d’Auzac & Jacob, 1989). The kinetics of single laticifer emptying and refilling have never been reported quantitatively, although qualitative observations on Euphorbia indicate that recovery to exuding pressure can occur within 1 min (Spencer, 1939); yet it must be important to plant wellbeing that, a reservoir for deterrent chemicals having been depleted but not destroyed by herbivory, it be expeditiously refilled. There seems to have been little if any experimental study of the turgor pressure of intact single laticifers; and the techniques of single-cell nano-osmometry or nano-manometry (e.g. Tomos & Leigh, 1999) have yet to be applied to laticifers.

Because the mature laticifer has not been reported to contain chloroplasts (e.g. Sacchetti et al., 1999) and is believed to have no functional plasmodesmatal connections with its neighbors, it seems reasonable to assume that it obtains its energy inputs from the apoplast and therefore that its electrophysiological behavior might display electrogenic sequelae of active transport. However, the electrophysiology of the laticifer is in its infancy; in fact, the few formal studies that exist are associated with F. Bouteau and collaborators (e.g. Bouteau et al., 1991, 1999a,b). In a nutshell, Hevea laticifers have a transplasmalemmal resting potential of c. −115 mV which is depolarized by 15–25 mV by the addition of glucose or sucrose to the bathing medium and is hyperpolarized by c. 40 mV by ethylene; seemingly possess an electrogenic H+-pump which activates sugar symporters; possess voltage-sensitive inward K+-channels; and yield protoplasts whose resting potential is only c. −35 mV, possibly as a result of overstimulation of mechanosensory channels.

Laticifer loading has been attributed to symplastic transport of nutrients from phloem to parenchymal cells adjacent to the laticifer, followed by entry into the apoplast, from which they are taken up into the laticifer by an H+-sugar symport system (Bouteau et al., 1999a) and turned into latex.

End-to-end transport in single intact laticifer cells has never been studied intensively. Laticifers, which putatively manufacture their own contents using resources received from phloem, which is nearby wherever laticifers are found, have no obvious need for long-distance mass transport in a normal sense. Indeed, there is as yet no body of firm evidence as to whether they are involved in long-distance transport of any sort or of any thing, even isoprenoids, although there are hints of possible long-distance transport in laticifers, both shoot-to-root (Santana et al., 2002) and root-to-shoot (St-Pierre et al., 1999). This statement excludes, of course, the obvious special case of rupture and subsequent drainage of the laticifer: this is not long-distance transport in the sense in which it is commonly used. Further, the existence or nonexistence of cytoplasmic streaming within laticifers seems never to have been systematically investigated. Moreover, experiments that failed to deliver evidence supporting manifest bulk flow would not thereby provide evidence against targeted transport mechanisms that selectively move specific biochemicals or organelles (trafficking).


Laticifers have been touted as a defense against herbivory for more than a century (Dussourd & Eisner, 1987). Nevertheless, heavily quantitative tests of this defense hypothesis have not been all that common, and a goodly percentage of the tests that were quantitative have been associated with David Dussourd and his collaborators (e.g. Dussourd & Eisner, 1987; Dussourd & Denno, 1991; Farrell et al., 1991; Dussourd, 1999, 2003). Typical of these tests that strongly support a defensive role for laticifers are the following.

  • • Insects that feed preferentially on a given laticiferous plant normally chew first through the veins that service the intended site of their herbivory or (alternatively) cut a trench around the leaf area that they intend to consume Dussourd & Eisner (1987). This strategy detaches the area distal to the cut and targeted for consumption from the latex stores proximal to the cut, and permits the latex within the targeted area to drain.
  • • Feeding insects that have their mouth parts artificially smeared with latex from the plant tend to stop feeding; to engage in mouth-cleaning activities; and to move somewhere else to resume feeding (Dussourd & Eisner, 1987).
  • • Experimental severing of veins in milkweed leaves increased the palatability of the distal portions of the leaves both to milkweed specialist insects and to herbivorous insects that normally avoid milkweed (Dussourd & Eisner, 1987).
  • • Insects that specialize in plants with branching nonarticulated laticifers typically are vein-cutters (Dussourd & Denno, 1991); vein-cutting effectively isolates a local area of leaf from the basal stores of latex proximal to the cut. Insects that specialize in plants with articulated anastomosing laticifers (which form a protective net throughout the leaf) typically are trench-cutters (Dussourd & Denno, 1991); and trenching is the only way of disconnecting the region selected for herbivory from these plants’ stores of latex. Insects that specialize in articulated nonanastomosing laticifer typically eat between the veins (Dussourd & Denno, 1991), and this seems a reasonable strategy when the latex ducts are independent of one another and their presence is not ubiquitous.
  • • Behavioral sabotage (vein-cutting or trench-cutting) can reduce the insect's ingestion of latex by upwards of 90% (Dussourd, 1999).
  • • Trenching behavior is not necessarily automatically practised by trench-cutting insects: it often must be elicited by chemicals encountered during feeding, and only specific substances within the latex are effective elicitors (Dussourd, 2003). Moreover, elicitation followed by cutting is not always a sure-fire recipe for success: the cutter must be able to tolerate the chemical cocktail ingested during cutting and also any residual chemicals in the drained laticifers then consumed (Dussourd, 2003).
  • • Generalist caterpillars that do not trench laticifer-containing leaves to inactivate the laticifers gain weight much more slowly on intact leaves than on leaves detached to inactivate laticifers (Dussourd & Denno, 1994).
  • • Aphids on inflorescences of lettuce (Lactuca sativa) have a markedly higher mortality than those on stalks and leaves because of entrapment in latex (Dussourd, 1995).
  • • Silkworm larvae thrive on leaves of mulberry (Morus spp.). Other caterpillars find them toxic, unless the latex containing alkaloidal sugar-mimic glycosidase inhibitors is first removed from the leaves; they also find normally wholesome diets toxic when mulberry latex is added (Konno et al., 2006).
  • • Chitinases have also been found in latices from various species (Graham & Sticklen, 1994). However, these are commonly thought to be directed against fungi rather than insects, a reminder that insects are not the only organisms that attack plants and against which defenses must be mounted.

In summation, there would seem to be a duel of lunge and parry between a laticiferous species and its enemies.

Open questions

The ecological significance of laticifers is, provisionally, understood: plants that have them deter insect herbivory and often are found unpalatable by foraging quadrupeds. To achieve these ends, laticifers normally contain sequestered chemicals that discourage their herbivores, and maintain internal pressures high enough to spew these chemicals onto or into herbivores that puncture them. Moreover, the chemicals in question are sometimes of considerable commercial importance. But there are still things it would be nice to know, a few of which are outlined below.

  • • The phyletic distribution of laticifers by type and by taxon has not yet been explained and might even be described as capricious. But wherever encountered they are commonly described as differing from their neighboring parenchymal cells in many ways besides their latices (e.g. Serpe et al., 2002).
  • They have been reported to exist in both the Fungi and the Plantae. Within the Fungi they are found in the family Russulaceae (e.g. Agerer 2006); so pronounced are the latices of genus Lactarius that they are mainstays in many classification schemes (e.g. and through it Within the Plantae they are known only in tracheophytes: in division Pteridophyta they are found in family Marsileaceae (e.g. Labouriau, 1952); possibly because members of this group are rich in resin ducts which serve a strongly analogous function, laticifers are said to be rare in division Coniferophyta, although they have been reported in the Gnetales (e.g. Tomlinson, 2003); in division Anthophyta they occur in many families of varying degrees of linkage (e.g. Farrell et al., 1991).

  • The release of latex upon being punctured seems not to have been a plant trait that has been systematically studied in its own right. Rather, it appears to have been a characteristic noted en passant as the organism was evaluated for other reasons; and in consequence, the taxonomic distribution of laticifers is poorly understood. This situation, however, might be remediable at minimal cost by a web-based Exudate Directory patterned after the less sharply focussed Wikipedia: after all, there are in aggregate a great many field botanists who could contribute reports of exudation without exorbitant effort. Verification of the precise nature and source of the exudate would require added effort; but a presumptive identification of an exudate as latex could be based upon simple microscopic examination of exudate, as only laticifer exudates are known to contain dense suspensions of particulate matter (cf. Esau, 1965).

  • • Simultaneous time courses of exudation from the two sections of a transected laticifer seem never to have been measured. Nor has the mechanism by which the exudation slows and stops been adequately investigated. These blanket statements must of course be qualified by citing the classical literature on rubber tree tapping (e.g. Bonner & Galston, 1947; Frey-Wyssling, 1952; Buttery & Boatman, 1976). For Hevea at least, falloff of exudation is monotonic, looks roughly exponential (cf. Frey-Wyssling, 1952), and has a final stoppage seemingly associated with the rupture of tonoplast vesicles called lutoids (cf. Buttery & Boatman, 1976; d’Auzac et al., 1995). This is but one species, is articulated-anastomosing, and contains lutoids: what of the hundreds of other genera?
  • • The static pressure within an intact laticifer has not been adequately studied, although this presumably would be possible employing modern pressure probe techniques (cf. Tomos & Leigh, 1999). One can define empirically at least three pressures within in a plant's tissue.

Local fluid pressure as measured at the tip of the micropipette of a pressure probe located in a gas or a liquid. Turgor pressure within a protoplasm is an example of this. This may be a relative pressure measured differentially with respect to the local atmospheric pressure as the micropipette is advanced into the tissue. However, its physiological importance normally depends upon its value compared with some nearby symplastic site or an apoplastic site, and therefore it should be converted to absolute pressure for reporting.

Tissue pressure as measured by an experiment (or Gedankenexperiment or numerical model) in which one determines the absolute pressure needed to inflate a suitable balloon inserted extracellularly into the tissue. This pressure seems clearly to be a property of an entire macroscopic region of minimally perturbed tissue.

Exudation pressure as measured by boring into the tissue and implanting a manometer which senses the regional apoplastic fluid pressure of an injured anatomy.

Exudation pressures have been measured for laticiferous tissue (e.g. Milburn & Ranasinghe, 1996), but the other two pressures have not. In particular, it is not known what the pressure differentials are between laticifer cells and their adjacent parenchyma.

  • • The dynamic turgor pressure variations deep within the two sections of a transected laticifer have never been measured. One might expect (1) a sudden drop in pressure associated with an elastic contraction of cell walls and an initial gush of latex (cf. Frey-Wyssling, 1952), (2) a slowing pressure drop as laticifer emptying continues because of osmotic water uptake from the apoplast (cf. Bonner & Galston, 1947, p. 476), (3) a bottoming-out of intra-laticifer pressure as latex gelation takes place at the cut and possible repair processes commence (cf. Buttery & Boatman, 1976, especially pp. 268–276), and (4) a resurgence of laticifer turgor pressure as osmotic equilibrium is re-established and the contents of the laticifer replenished. Despite the importance to the plant of repair and refilling of its defense system, evidence for step (4) rests largely upon a single long-forgotten brief communication (Spencer, 1939).
  • • If repair and recovery do indeed take place, they will presumably involve electrogenic processes within the laticifer. However, nothing is known of the comparative electrophysiology of the laticifer and of its parenchyma. Obviously, measurements that might be taken with a modern cell pressure probe as adapted for simultaneous measurements of pressure and voltage (i.e. Wegner & Zimmermann, 1998) would be more valuable than either potential or pressure measurements alone.
  • • Long-distance transport in the absence of damage and drainage has yet to be been ruled out for the laticifer. Neither cytoplasmic streaming nor trafficking has ever been investigated for a laticifer – any laticifer. It ought to be possible to study this qualitatively by microinjecting a nonreactive fluorescent label, following its spread with an optical sectioning microscope, and comparing these data with those expected from diffusion.



Secretory cells, secretory canals/ducts, and secretory glands are widely distributed in the tracheophytes (e.g. Fahn, 1979, 1988). Although their contents of gums (e.g. mucilage), essential oils (e.g. peppermint flavoring), and various resins (e.g. pitch) are familiar to almost everyone4, the literature dealing with issues of interest here is somewhat sparse. One can find reviews that have developmental/morphological focus (e.g. Nair, 1995) and reviews of defensive resin biosynthesis (e.g. Trapp & Croteau, 2001). In addition, the atlas of micrographs by Svoboda et al. (2000), although of restricted coverage, is superb as far as it goes. Detailed studies have been confined largely to commercially important species. Comparative information about taxa with secretory ducts seems localized to an appreciation of their wide distribution and to their commercial importance (e.g. Nair, 1995), but encyclopedic surveys of either are rare to nonexistent. Recent research has focussed on the conifers and has been largely biochemical or ecological, and it appears that the only biophysical intraluminal study (whether by pressure probe or by microelectrode) is that of Aung et al. (2001), who suggested that the 0.35 ± 0.14 MPa pressure within the oil cavities of lemon (Citrus limon) exocarp could result from the partial pressure of volatiles in the oil.

Anatomy and development

Secretory ducts, like laticifers, tend to be associated with vascular bundles, and it has been strongly argued that ‘the two are so functionally similar that they can be considered to constitute a single defensive syndrome despite differences in their anatomy’ (Farrell et al., 1991, p. 883). Nevertheless, the duct, unlike the laticifer, is extracellular5 as a result of its development by schizogeny (i.e. a spreading apart) or, possibly, lysigeny (i.e. programmed death and dissolution) of cells along its route (Nair, 1995; Turner, 1999); and normally it is lined with secretory cells, termed an ‘epithelium’ by Esau (1965, p. 254), Fahn (1979, Ch. 9), and Nair (1995). Schizogenously derived ducts originate from the dissolution of the middle lamella between the duct initials and the formation of an intercellular space (Nair, 1995). The duct lumen forms subsequently by anticlinal division of the initials, anticlinal flattening of the cells that bound the lumen, and development of these peripheral cells into characteristic epithelial cells which are the major sites of synthesis and secretion of the natural products that fill the duct lumen. The ducts are normally oriented parallel to the longitudinal axis of the organ containing them, but they may anastomose tangentially.

Secretion into the lumen may proceed either by transplasmalemmal secretion on a molecule-by-molecule basis, or by vesicular exocytosis, or by both (merocrine secretion); occasionally it may proceed by lysis of epithelial cells with direct release of cellular contents into the lumen (Nair, 1995).

General physiology

The study of the physiology of a secretory duct is complicated by the fact that its contents are extracellular, nonliving, and dependent upon its epithelium for maintenance moreover, they are not necessarily aqueous but frequently viscous, sticky, and capable of promoting apparatus malfunction either electrically or mechanically. However, the pressure that characterizes regions of tissue was, in the distant past, often measured and was normally called ‘exudation pressure’ or ‘secretion pressure’ (cf. Stark, 1965). Such measurements (e.g. Stark, 1965, p. 313; Milburn & Ranasinghe, 1996) are commonly made by drilling or punching a small hole into the plant to a depth at which ducts should have been transected and then inserting into the hole a slightly larger steel tube, one end of which is attached to a pressure gauge. When, some 30 yr ago, the author witnessed such a device in use at the Rubber Research Institute of Malaysia's facility outside Kuala Lumpur, he considered the technique to be mechanically robust and capable of being mastered easily. Unfortunately, the tubes used are normally several hundred micrometers in outer diameter, and the variety of structures making contributions to the measured pressure is uncertain. That said, exudation pressures of several hundred kilopascals are normally recorded.

It seems to be conventional wisdom (Stark, 1965, p. 314) that ‘resin exudation is largely regulated by the osmotic pressure of the epithelial cells lining the resin ducts’. But, if the resin is a hydrophobic liquid, it is not obvious why the osmotic properties of the cell interior are directly relevant; that is, osmosis (as commonly conceived of) describes water transport between two aqueous phases, whereas liquid resin and water are immiscible. Nevertheless, in all cases, the osmolality of epithelial cell interiors must presumably be large enough to prevent their being collapsed by a high duct pressure which would express symplastic water through their plasmalemmal aquaporins in a process of reverse osmosis. Finally, only if resin egress from the duct is mechanically restricted should exudation be a function of duct pressure; otherwise exudation ought to be principally a function of the rate of merocrine secretion that the duct epithelium can maintain (cf. Nair, 1995, p. 333).


As emphasized previously, it is well accepted that a principal function of secretory cells, ducts, and glands is to deter herbivory (Farrell et al., 1991), especially in the Pinales (Francheschi et al., 2005), and it is a commonplace that the aversive or toxic effects of a plant's secretions are rather more severe upon animals that seldom eat it than upon its principal attackers. It now appears that this balance between plant defense and herbivore attack is more than just an isolated evolutionary arms race: although, in the absence of a broad and well-accepted theory of plant defense, how much more is problematic (cf. Stamp, 2003). What seems unassailable is that putatively noxious secretions are important to the plant's wellbeing, although only isolated aspects of this importance have been explored in great detail: they occur, after all, not only as contents of laticifers and secretory ducts, but also in single cells or compact groups of cells, as in citrus oil glands or Cannabis leaves (cf. Kim & Mahlberg, 1997).

Therefore, from an ecological viewpoint, an issue of significance is the question of what might be changed about the overall exudate system of a particular species to make it a more effective deterrent to herbivory. This question seems especially relevant in an era when the capabilities of herbivore and plant need no longer be tightly coupled by coevolution because that linkage could, hypothetically, be vastly weakened by genetically re-engineering the plant (cf. McCaskill & Croteau, 1999); that is, rather than having two species that evolve together incrementally in an endless sequence of minute mutations, one or the other could hypothetically be given a sudden huge comparative advantage. However, the difficulties of such a program are commonly expected to be great, even for so intensively studied a sub-area as isoprenoid biosyntheses. If ‘merely’ tweaking-to-specification the mixture of volatiles in the exudate has been found challenging, it seems likely that such strategies as dynamically controlling synthesis rate or storage pressure of the exudate will be prove daunting – as, by modern standards, neither observable has been studied to any significant degree.

Open questions

Because the laticifer and the secretory duct have many of the same functions (e.g. Farrell et al., 1991), the introductory paragraph and nearly all of the questions following it in the ‘Open questions’ part of the ‘Laticifers’ section will be found germane. Additional questions of particular relevance to ducts are the following.

  • • As resin is thought to be a mixture of isoprenoids and not an electrolyte solution with a possible dispersion of hydrophobic particles, it presumably does not have a well-defined electric potential. However, a gum is hydrophilic and a gum duct could well contain an electrolyte solution. Thus, all of the micropipette physiology suggested for laticifers should be relevant to gum ducts. However, micropipette investigations will not necessarily be easy to carry out as both gums and resins tend to be viscous/sticky and could well clog the tip of the micropipette.
  • • The phenomenology of exudation from a resin duct might be qualitatively different from that of a laticifer because the laticifer might be able to refill itself by providing primarily solute, the solvent being strongly supplemented by osmotic uptake. By contrast, virtually everything in the resin duct is presumed to be secreted by its epithelial cells with no boost from osmosis.
  • • The extensive and valuable behavioral studies of caterpillars interacting with laticifers (e.g. Dussourd, 2003) seem not to have been repeated for bark beetles interacting with resin ducts.
  • • A key to the bark beetle's success is its ability to avoid deluge by resin while penetrating a tree's periderm to reach the phloem/cambium zone, where it feeds and breeds. This it accomplishes by burrowing through the spongy tissue of a lenticel where the density of resin ducts is low and the tissue itself less tough mechanically (cf. Rosner & Führer, 2002, especially fig. 1). The ability or willingness of the beetle to burrow through various kinds of tissue has never been quantified, but, when it has, operations research-type modeling of the local anatomy might suggest a better mix of air space, resin defense, and secondary cell wall toward which to breed.
  • • The presumed importance of resin ducts in the defense of valuable timber crops is well known (e.g. Stark, 1965; Trapp & Croteau, 2001; Francheschi et al., 2005). Nonetheless, over wide areas of the globe, infestations of herbivorous beetles pose an unpredictable threat (cf. of the United States Department of Agriculture Forest Service). The strategies of defending against such attack have been recently reviewed and the paucity of immediately applicable tactics highlighted (Mumm & Hilker, 2006). Because the channeling of metabolic activity into chemical defense is costly to the plant and diverts resources that otherwise might be devoted to maintaining robust health, added research emphasis might usefully be given to the following.

A search for chemicals that, when exogenously applied to a stand, reliably upregulate inducible defenses. If a single plant senses an insect infestation and then upregulates, the upregulation could come too late. Presumably, however, the forester could anticipate the threat and place an entire stand on alert.

Identification of the chemical aggregation signals that enable the insects to swarm a single tree and overcome its constitutive defenses by weight of numbers. Given a thorough understanding of the control of aggregation, it should be feasible to deploy either decoys (possibly lethal) or chemical ‘static’ which swamps any genuine aggregation signals.

Discovery of chemicals that a problem herbivore finds aversive. A spray that inhibits aggregation could be a major help.

Magic bullets are not to be expected, although the sheer number of isoprenoids currently known might tempt one to believe that some quite potent candidates could exist for the applications described above; recent review articles relevant to these thrusts include Francheschi et al. (2005) and Cook et al. (2007). In summary, the evolutionary balance between successful defense and triumphant attack appears delicately poised, and the availability of new interventions (even minor ones) might let a forester tip the balance markedly in favor of the trees.


In engineering a plant to achieve some specified goal, it does not suffice merely to localize a property anatomically: one must understand also structure, function, and the inevitable unexpected linkages that appear when one tries to tweak a single parameter. Some of the many complexities of the bioengineering challenges associated with secretory structures have been discussed, for example by McCaskill & Croteau (1999), and the latter also emphasized the paucity of information available on the physiology of these structures. Nevertheless, when enough preliminary studies have been completed, it is possible to obtain promising results; for example, the much studied peppermint plant (Mentha×piperita) has been successfully engineered to improve both the composition and the yield of its essential oil. However, this was achieved only after decades of painstaking toil climbing the learning curve; one can get a flavor for this by examining the ‘Croteau Festschrift’ (Phytochemistry 67, issues 15 and 16, 2006) to understand the magnitude of the activity in terpenoid biochemistry that underpins this achievement.

The challenge, for both laticifer and duct, is that too little of this basic orienting spade work is being done. And, in comparison to phloem and xylem, their physiology is a terra incognita and probably will remain so until this trend is reversed. Of course, classical plant breeding throve by painstaking selection to optimize complex but desirable endpoints; and it did so without being overly concerned with the anatomical and biochemical modifications that underpinned the improvements. But there are limits to such a strategy, two of which are that progress can be painfully slow and that one usually arrives only at a local optimum and stalls there. For a set of traits whose involvement with a desirable endpoint seems evident, a different strategy is to believe that it is only common sense to study those traits in depth. One does not have to explain precisely how a deeper understanding of one or another of the traits will help: because – until one thoroughly understands the desirable endpoint – one cannot know which items of basic understanding proved to be the most valuable. This truism certainly proved correct in the development of efficacious sources of artificial illumination, and it also has played out over many decades in the steady improvement of motor cars.

Basic research, even well guided, is time consuming. It is also costly; and, in retrospect, much of the money spent will turn out to have been money wasted. However, if we desire rationally to design improved plant defense systems, do we have a choice? Absent a vast store of information with which to reason, rational design is a pipe dream.