It is proposed here to use both molecular and morphological data sets. To characterize their phylogenetic signal the character state matrices should be analyzed separately, using parsimony (morphology) and Bayesian methods (morphological as well as molecular). In addition, molecular and morphological data sets should be combined into one joint analysis. With respect to morphological and molecular characters two approaches may be followed: (1) morphology is merely optimized in post-tree analysis of the molecular results, or (2) morphological and molecular characters are combined into one data matrix. The first approach is favoured in many recent studies and considered to be the only contribution of morphology to phylogenetic analysis by Scotland et al. (2003); but see Jenner (2004) for a rebuttal. It is here suggested that the second approach, i.e., combined analysis, is applied.

From an empirical perspective, it has been shown that morphology can have a profound effect on the combined analysis, irrespective of the fact that the number of molecular characters generally exceeds the number of morphological features (Jenner 2004; Assis 2009). Combined analyses may be different from the analyses of the separate data matrices and consensus trees may hide a phylogenetic signal that is generated by a total evidence analysis. Positive contributions of morphology to quantitative clade support measures in combined analyses have been observed for a large number of taxa (Jenner 2004).

As the fossil record of flatworms is sparse and does not provide adequate calibration points for a molecular clock, calibration of the phylogenetic timetrees has to be based on other kinds of data, such as, for example, paleogeographical information (see above: Impact and innovative aspects). This means, for example, that we will be looking for closely related taxa a and b that are endemic to the areas A and B, respectively. In addition, we will be looking for those areas A and B inhabited by endemic taxa for which paleogeographic data indicate the time since the two areas have fragmented from a single ancestral area. These two pieces of information, together with the molecular clock hypothesis, will enable one to date all cladogenetic and biogeographic events in the entire lineage of which a and b only form a part.

Within the triclad flatworms there is a good number of such disjunctions, due to vicariance events, within species or between closely related species, that may be tested as possible calibration points. For example, land planarians of the genus Othelosoma Gray, 1869 are restricted to Africa and India and have attained their current distribution when India and Africa started to separate at about 150 Mya. Table 1 specifies the taxa, and their presumed vicariance events and divergence times that may profitably be used for calibration.

One may perhaps be inclined to consider paleogeographical calibrations to be less ideal than fossil calibrations. In point of fact, the opposite may be the case. Fossil-calibrated molecular clocks at best provide minimum dates (Wilke et al. 2009). Therefore, it has been argued that clocks are best calibrated with reference to the distribution of molecular clades and associated tectonics (e.g., Azuma et al. 2008; Heads 2014).

Molecular dating is a rapidly developing field and therefore there is currently no single best method; each approach has its advantages and disadvantages. It is also important to note that in the phylogenetic tree of the Platyhelminthes, the triclads constitute one of the crown groups (cf. Baguñà and Riutort 2004). This implies that the nodes below the Tricladida represent older taxa that therefore might be used as potential outgroups in the phylogenetic analysis.

The molecular data matrix for this project may be derived from published data from GenBank, and new molecular sequences generated from fresh material of new taxa examined during the project. Previous studies have shown that the following genes provide the best resolution at the hierarchical levels of the phylogenetic tree of the triclads that form the focus of the putative research programme: nuclear 18S rDNA and 28S rDNA for resolving the deeper, more ancient branches in the tree, and mitochondrial COI to contribute signal on more recent splits (e.g. between and within genera) (Álvarez-Presas et al. 2008). Subsequently, another nuclear marker, the elongation factor 1-alpha (EF), was added to the list of informative markers for an intermediate level between family and species, contributing to resolution of the relationships between members of the Geoplaninae subfamily (Carbayo et al. 2013).

It has been suggested that many of the problems associated with the amplification and sequencing of planarian molecular markers may be solved by applying next generation sequencing (NGS) methodologies (Sluys and Riutort 2018). This new era has started already, and in recent years genomes and transcriptomes of freshwater planarians have already been sequenced (Egger et al. 2015; An et al. 2018; Grohme et al. 2018; Rozanski et al. 2018), contributing new anonymous nuclear markers that can be used as phylogenetic tools. There are even bioinformatic pipelines that can help in the discovery of these markers using NGS data as input (Frías-López et al. 2016). This has already been applied in freshwater planarians, and will soon be used also for terrestrial flatworms. Moreover, there is now available also information on other mitochondrial markers, in addition to COI, as whole mitogenomes were sequenced and annotated (Sakai and Sakaizumi 2012; Solà et al. 2015; Gastineau et al. 2019; Yang et al. 2019), contributing to the combination of nuclear and mitochondrial markers, thus obtaining better resolved phylogenies.

The project requires a large morphological data matrix of all major taxa of land planarians (e.g., for the current 55 genera; cf. Sluys et al. 2009). For this, the programme may be taking as leads the doctoral dissertations of both Dr L. Winsor (2003, James Cook University, Townsville) and Dr F. Carbayo (2003, University of Sao Paulo). Both researchers made detailed character codings and scorings for taxa of land planarians. Dr Winsor mostly focussed on Australian taxa, while Dr Carbayo concentrated on South American species, although neither used these scorings for a subsequent phylogenetic analysis. During the project character codings and scorings may be extracted from these two doctoral theses and, subsequently, be combined, refined and also supplemented by scorings for taxa not examined by Winsor or Carbayo. Definition and scoring of morphological character states should comply with the most recent insights (cf. Sereno 2007).

The morphological, molecular, and combined data matrix may be analyzed using maximum parsimony (e.g., TNT; Goloboff et al. 2003), maximum likelihood (e.g., RaxML; Stamatakis 2014; IQtree; Nguyen et al. 2015), and Bayesian Inference (MrBAYES, Ronquist et al. 2012, and BEAST, Drummond et al. 2012). Informative genes should first be analyzed separately, but also combined in the form of concatenated datasets. Substitution model parameters may either be selected using nested likelihood ratio tests or explored using reversible jump Markov chain Monte Carlo methods (jModeltest, Darriba et al. 2012; Partitionfinder, Lanfear et al. 2017). Saturation tests should be performed using DAMBE (Xia 2017).

Molecular data may be analyzed also under a dynamic approach to homology. In the dynamic approach delineations are dependent upon the topology of the phylogenetic trees on which they are optimized. In this context hypotheses of homology are part of phylogenetic hypotheses and are subject to the same optimality criteria as the trees, viz., minimisation of evolutionary transformation events. The computer program POY (ver. 4.0 beta 2635; Wheeler et al. 2006, Varón et al. 2010) allows for phylogenetic tree searches under this dynamic homology or direct optimization approach.

Aquatic planarians should form one of the outgroups for the phylogenetic analysis. Particularly the freshwater family Dugesiidae has been shown to share a close relationship with the land flatworms (see Sluys et al. 2009). But other freshwater groups, as well as marine triclads, should be included also as outgroups. Comparison with these freshwater and marine forms will reveal the adaptations that made possible the transition to land.

The research proposed here seeks to explore the evolution of terrestrialization of land flatworms in time and in correlation with presumed key morphological adaptations. As regards the time axis, this will be important both for direct reconstructions of when terrestrialization happened, as well as in the subsequent analysis of whether hypothesized key innovations co-vary in their location on the phylogeny with elevated diversification rates under a model of adaptive radiation (e.g., using the method of Ree 2005). It is here proposed that divergence dates estimates may be reconstructed from molecular data calibrated using previously published node ages (Hedges and Kumar 2009) in a relaxed molecular clock approach as implemented in BEAST (Drummond et al. 2012). As for covariance between traits (see above: Key innovations, adaptive radiation and terrestrialization), the general model will be that the radiation of flatworms, with presumably repeated terrestrializations, constitutes a natural experiment such that morphological changes hypothesized to be associated with terrestrialization co-vary with and co-occur in ancestral state reconstructions on a phylogenetic tree. Sophisticated comparative methods to correct for autocorrelation (because the different lineages are related to each other) have been developed, including Bayesian statistical methods that allow for phylogenetic uncertainty (e.g., Pagel and Meade 2006). Understanding of the evolutionary dynamics and consequences of such innovations will contribute to the unraveling of their genetic basis and their role in speciation events and, consequently, the adaptive radiation of this group of animals.

To give an example, the study will provide insight into the evolution of creeping soles, the latter defined as: “A flat or ridged modified strip of epithelium on the ventral surface of geoplanid triclad flatworms characterized by the presence of cilia …. which provides propulsive forces by ciliary or muscular action, or by a combination of both” (Sluys et al. 2009: 1773). Traditionally, a narrow creeping sole is considered as the primitive character state, while medium to broad soles and loss of a creeping sole are seen as derived states. However, narrow soles may also secondarily result from partial reduction, while absence of a creeping sole altogether may be the ancestral state, as aquatic forms lack one. However, the evolution of creeping soles has never been analyzed within a phylogenetic context; the research programme developed here will be the first to do so. If the traditional view holds true, one would expect to infer the narrow creeping sole as the character state at the root of the tree, with transitions to medium or broad soles, or loss of a creeping sole, nearer the tips. Under the opposite scenario, the absence of a creeping sole being the ancestral state, one would consequently expect this state to be reconstructed for the root of the phylogeny with alternately polarized state changes reconstructed nearer the tips. Likewise, expectations can also be formulated for body shape: if cylindrical body shapes indeed conserve water one would expect transitions to this morphology to coincide with, or follow terrestrialization when reconstructed on the phylogeny. In a similar way, all of the hypotheses mentioned above (see above: Key innovations, adaptive radiation and terrestrialization), as well as others that may arise during the study, can be tested.

Introduction

A particular crucial feature that may be among the most difficult to examine and to plot its character states on the phylogenetic trees concerns the adaptation to terrestrial respiration. Planarian flatworms possess neither circulatory nor respiratory systems for transporting oxygen or digested food substances to the internal tissues. In these animals, oxygen is absorbed across the entire body wall and for this diffusion process water is required to dissolve oxygen and carbon dioxide in order to cross cell membranes. Clearly, this poses no problem for freshwater and marine planarians as they live in an aquatic habitat, but when ancestral planarians colonized the land, leaving this aquatic milieu must have formed a major hurdle, as the physical properties of water and air are so different. It is true that many land planarians live in habitats with a high humidity, but still these conditions greatly differ from a fully aquatic environment, while there are also terrestrial planarians that occur in mesophile and xerophile habitats (Froehlich 1955; Winsor et al. 1998).

Thickness of flatworms

The analysis of this subject is complicated by the fact that not much is known about the respiratory physiology of free-living flatworms in general and planarians in particular. Furthermore, most of these studies concern aquatic species. A striking example is McNeill Alexander’s (1979) conclusion that the maximum possible thickness of a free-living flatworm is 1.0 mm or at most 1.5 mm. His conclusion is based on an argument that involves the calculation of the diffusion rate of oxygen. In philosophical context McNeill Alexander’s type of functional explanation has been presented as an example of viability explanations, which are distinct from causal or historical explanations (Wouters 1995).

McNeill Alexander’s (1979) calculation is based on a number of factors, viz., (a) partial pressure of oxygen in water, (b) rate of gas diffusion, (c) the fact that oxygen mostly will diffuse through the dorsal body surface, as the ventral surface is in contact with the substrate, (d) the diffusion constant for oxygen diffusing through connective tissue, (e) density of flatworm tissue. When the known or estimated values for these variables are used in a derivation using Fick’s law of diffusion, “This calculation indicates that the maximum possible thickness for a flatworm … is about 0.5 mm if oxygen diffuses in only from the dorsal surface, or 1.0 mm if it diffuses equally from the ventral surface…. A similar calculation for a cylindrical turbellarian indicates that the maximum possible diameter would be 1.5 mm …. This is probably why large flatworms are flat.” (McNeill Alexander 1979: 185).

As flatworms rely entirely on oxygen diffusion through the surface, it is indeed advantageous to have a large surface area:volume ratio, i.e., to be flat and not cylindrical. In general, this holds true for aquatic species, while within species their thickness hardly or not at all increases with an increase in plan area of the body (Calow 1987). Another interpretation of the broad bodies of large land flatworms is that this facilitates capture and subdual of prey (cf. Cseh et al. 2017), i.e., forms an evolutionary adaptation to new kinds of terrestrial prey not encountered by their aquatic ancestors.

For an organism that relies on direct diffusion of oxygen Prosser (1973) provided a formula for the calculation of the thickness of the animal, based on the oxygen concentration of the medium, a diffusion coefficient K, and the rate of oxygen consumption. When the calculation is done for both water and air (Table 2) it reveals that air-breathers can be six times thicker because there is more oxygen in air than in water (it should be noted that there is uncertainty about the units in which thickness is expressed).

Nevertheless, the conclusion of McNeill Alexander (1979), as well as the calculation on which it is based, is open to a number of criticisms. First of all, it is certainly not the case that large flatworms are always very flat or thin. For example, “giant” freshwater species from Lake Baikal, such as Bdellocephala bathyalis Timoshkin & Porfirjeva, 1989, are certainly thicker than 0.5–1.0 mm (B. bathyalis measuring 3.3 mm in thickness in preserved condition; Sluys et al. 1998). And also the large land planarians are usually thicker, such as Polycladus gayi Blanchard, 1845 (4 mm thick), Pseudogeoplana lumbricoides (Schirch, 1929) (3.5 mm), P. nigrofusca (Darwin, 1844) (3 mm), Geoplana rufiventris Schultze & Müller, 1857 (2.3 mm) (Von Graff 1899). Although such measurements on the thickness are usually made on preserved specimens, and that thus fully stretched live specimens will be thinner, this does not necessarily compromise the picture. It should be realized that most of the time the flatworms hide under stones or fallen logs, etc. and that during this resting period they are highly contracted, thus presumably approaching the condition of preserved specimens.

Another example of a thick flatworm-like animal is the basal bilaterian Xenoturbella bocki Westblad, 1950, which may reach a length of 2–3 cm and a thickness of 5 mm (Franzén and Afzelius 1987).

Another objection that may be raised against McNeill Alexander’s (1979) calculation concerns his assumption that diffusion in flatworms is more or less equal to that in frog muscle and connective tissue, i.e., 2 × 10⁻⁵mm²atm⁻¹s⁻¹. However, it is doubtful that the mesenchyme and gut tissue—forming the major component of the planarian body—have the same diffusion constant as frog muscles and connective tissue. Evidently, the diffusion constant of planarian tissue will not be equal to that in water (6 × 10⁻⁵mm²atm⁻¹s⁻¹) but perhaps 4 × 10⁻⁵mm²atm⁻¹s⁻¹ would be a value that is more realistic for flatworms. This would then imply a higher rate of diffusion and, consequently, would allow the worms to be thicker.

For the rate (m) of oxygen consumption per unit volume of tissue, McNeill Alexander (1979: 185) used 0.1 cm³ oxygen g⁻¹h⁻¹, “or a little more”. However, according to Hyman (1951: 207 and references therein) this may be higher, viz., 0.2–0.31 cm³ oxygen g⁻¹h⁻¹, that is “0.2–0.3 cc per gram per hour in adult worms.” According to Moore (2006) it would be about 0.1–0.2 ml oxygen g⁻¹h⁻¹ at 15 °C.

A variable that is not taken into account by McNeill Alexander (1979) is the relationship between size and metabolic rate in flatworms. Although data are scant, generally respiratory rate decreases as size increases (Vernberg 1968).

Although McNeill Alexander’s view on the maximum thickness of flatworm-like animals has been explicitly or implicitly endorsed (e.g., Ruppert et al. 2004; Moore 2006), he considered it merely “a very rough calculation in a textbook” and he conceded that “There are many doubtful assumptions in the calculation” (McNeill Alexander in litt., 22 August 1996).

Respiratory pigments

Apart from the doubtful assumptions in McNeill Alexander’s calculation, another explanation for the empirical fact that flatworms frequently are thicker than 0.5–1.0 mm may lie in the presence of respiratory pigments, transporting oxygen across tissues. Hemoglobin is the most common respiratory pigment among invertebrates in general (Brusca and Brusca 2003). With respect to flatworms it has been reported mostly from parasitic helminths (Weber and Vinogradov 2001) as well as for a number of ecto- and endosymbiotic turbellarians, such as Paravortex Wahl, 1906, Triloborhynchus Bashiruddin & Karling, 1970, Cleistogamia Faust, 1924, Serita Cannon, 1982, Paranotothrix Cannon, 1982 (Jennings and Cannon 1987; Jennings 1988, 1997). Hemoglobin is known also from the free-living flatworms, viz., a species of Phaenocora Ehrenberg, 1835 (Vernberg 1968; Weber and Vinogradov 2001). It has been hypothesized that presence of hemoglobin in the entosymbiotic graffilid rhabdocoel Paravortex scrobiculariae (Graff, 1882) represents an adaptation to the oxygen-poor conditions to which this species is subjected, in contrast to the well-aerated habitats of other species, such as P. cardii (Hallez, 1908) and Graffilla buccinicola Jameson, 1897, which lack hemoglobin (Jennings 1981). Unfortunately, oxygen carrying components such as hemoglobin have not been studied in planarians. Winsor (1998) mentioned the possible respiratory role of uroporphyrins (other than hemoglobin) present in the rhabdoids of Platydemus manokwari de Beauchamp, 1962. Potential presence of hemoglobin may be deduced from the presence of non-epidermal pigments, often situated in or around specific organs, such as the brain and the pharynx (Jennings 1981, 1988; Jennings and Cannon 1987). In planarians, body colouration generally is due to the presence of granular pigments located in the mesenchyme directly below the dorsal epidermis (Sluys and Riutort 2018), while in most species of the freshwater genus Girardia Ball, 1974 the pharynx is also pigmented. Therefore, red pigment at other locations in the planarian body may point to the presence of hemoglobin. However, I have rarely observed pigment to be present at such other locations, at least in histological preparations, suggesting that hemoglobin is absent in triclad flatworms.

Respiration in triclads and other free-living turbellarians

Oxygen consumption in planarians and other free-living turbellarians has been studied chiefly in freshwater species (cf. Vernberg 1968; Heitkamp 1979) and only once in a terrestrial species (Daly and Matthews 1982). An important result that emerged from these studies is the relationship between body size and respiration, in that generally smaller animals have a higher rate of oxygen consumption than larger specimens, when determined on a weight-specific basis; however, some species showed the reverse correlation, while for others no relationship could be established between body size and oxygen uptake (Vernberg 1968 and references therein; Heitkamp 1979).

Rates of oxygen consumption may be determined and expressed in different ways and are generally influenced by the temperature of the habitat. The type of response to changes in temperature varies per species, as some species are eurythermal and others much more stenothermal in their ecological requirements (Vernberg 1968 and references therein). Therefore, respiration rate frequently is expressed as microliters-O₂-per-(milli-)gram-wet weight-per hour (µlO₂/(m)g/WW/h) at a particular temperature. Another way to express oxygen consumption is by calculating the coefficient b, or the regression slope of a particular power function for the relation between body size and respiration. Unfortunately, the value of b varies much as it greatly depends on the size of the animals, their physiological condition, and many external factors (Heitkamp 1979).

For five species of freshwater planarians (Dugesia gonocephala (Dugès, 1830), Crenobia alpina (Dana, 1766), Polycelis nigra (Müller, 1774), P. felina (Dalyell, 1814), Schmidtea polychroa (Schmidt, 1861)) the following values were found for oxygen consumption (µlO₂/g/WW/h), measured at a temperature of about 15 °C: 170, 240, 135, 199, 116, respectively (Vernberg 1968). Unfortunately, these values for triclad flatworms cannot be compared directly with those determined for various microturbellarians as these were calculated per milligram wet weight (µlO₂/mg/WW/h), resulting in the following values, measured at a temperature of 15 °C: 1.242 (Dalyellia viridis (Shaw, 1791)), 0.688 (Opistomum pallidum Schmidt, 1848), 0168 (Mesostoma ehrenbergi (Focke, 1836)), ranging between 0.298 and 0.700 (eight different populations of Mesostoma lingua (Abildgaard, 1789)) (Heitkamp 1979).

The habitat temperature of the tropical terrestrial land planarian Bipalium kewense Moseley, 1878 is usually much higher than that of the aquatic triclads mentioned above, albeit that this invasive species has established itself outdoors in, for example, several North American states, the West Indies, Portugal, French Guiana, and France (Sluys 2016; Justine et al. 2018). The oxygen consumption of B. kewense specimens from outdoor localities in Arkansas, USA was determined at temperatures varying between 27–33 °C, which yielded respiration rates (µlO₂/g/WW/h) ranging between 113–290 (Daly and Matthews 1982). It is noteworthy that these values are in the same order of magnitude as those determined for freshwater planarians.

The coefficient b based on oxygen consumption of entire specimens of B. kewense ranged between 0.686–0.753, as measured at temperatures ranging between 27–33 °C (Daly and Matthews 1982). These values are in the same order of magnitude as those determined for the freshwater triclads Crenobia alpina (0.66), Dugesia gonocephala (0.82), and Polycelis felina (0.82) and the microturbellarians Mesostoma ehrenbergi (0.625), M. lingua (0.850), and Opistomum pallidum (0.880) (Heitkamp 1979 and references therein).

Terrestrial respiration

Terrestrial flatworms face two problems that involve mutually conflicting adaptations, viz., desiccation and respiration. A cylindrical body, with less surface area:volume ratio, will minimize water loss but restricts diffusion of oxygen to the internal tissues. Probably this is the reason why smaller terrestrial species tend to be round or oval in cross-section, e.g., species of the land planarian genus Microplana Vejdovsky, 1890. Therefore, large species tend to be flattened to create a large surface area:volume ratio in order to facilitate diffusion of oxygen to the deep tissues. Nevertheless, the generally large and particularly long species of the land planarian subfamily Bipaliinae Von Graff, 1896 also have a more or less cylindroid body in cross-section.

The partial pressure of oxygen in well-aerated water in equilibrium with air is 0.21 atm (McNeill Alexander 1979) and results in about 9 mgO₂/l at 20 °C. Evidently, the actual amount of dissolved oxygen available at a particular aquatic habitat depends on the temperature, depth, altitude, and the mixing properties of the water (e.g., running water in shallow streams mixing better with air). The amount of oxygen in air is about 30 times that of water (Little 1990; Moore 2006), amounting to about 210 ml O₂, weighing 280 mg, in 1 L of air (Schmidt-Nielsen 1975), while the diffusion rate of oxygen through air is much faster than through water (3 million times faster through air; Prosser 1973).

In the present context it suffices to realize that, thus, availability of sufficient oxygen would not have formed a stumbling-block during an evolutionary transition from water to a terrestrial environment. But, clearly, the organisms needed to evolve respiratory adaptations enabling them to extract oxygen from the air, as opposed to their ancestors, which had evolved in an aquatic habitat.

One such adaptation may be the production of mucus, which is secreted by both aquatic and terrestrial planarians. Mucus plays several important roles in the life of a planarian flatworm and is produced by various kinds of gland. Secretions from glands at the body margin produce a slime trail that facilitates the gliding movement of both aquatic and terrestrial triclads, effectuated by the propulsive force of cilia on the ventral body surface (Jones 1978). Many species of land planarians do not have such dedicated marginal glands, and various types of mucus are secreted from their creeping sole and/or ventral surface. Sticky mucus discharged by cephalic glands partakes in the capture of prey and has neurotoxic properties (cf. Thielicke and Sluys 2019 and references therein). The mucus of triclad flatworms also has repellent properties as the worms are only rarely eaten by other animals since their surface secretions appear to have repugnatorial function (Hyman 1951; Winsor 1998).

As mucus is produced also in aquatic triclads it could well be that these various functions already formed evolutionary preadaptations of similar functions in land planarians. Perhaps the secretion produced by the marginal adhesive zone is an exception as it has been suggested that it may not provide a lubricant for locomotion in the land planarians but form a moisture-retaining sealant in a resting animal (Winsor 1998). However, one function does not exclude the other.

Production of surface secretions may also have formed a preadaptation for respiration in the terrestrial environment as it covers the body with a “watery” layer that presumably improves the uptake of oxygen. This may be related to the possible role, including respiration, of substances (porphyrins) in rhabdoids that are conspicuous in the dorsal/dorso-lateral epithelium (and microrhabdites over the ventral surface) of land planarians (see above). To the best of my knowledge, this aspect of mucus secretion and respiration in land flatworms has never received any attention.

Evidently, uptake of oxygen through the body wall is only the first step in the respiration process. Hereafter, the oxygen needs to be transported to tissues deeper inside the planarian body. In the absence of respiratory pigments (see above: Respiratory pigments) this can be achieved only by means of diffusion. This implies that also the internal tissues must have a sufficient amount of water in order to be able to dissolve oxygen and carbon dioxide. Maintaining a sufficient level of hydration may be unproblematic for freshwater planarians but may require certain adaptations in marine and terrestrial forms. In particular, land planarians have no physiological or anatomical adaptations for water retention (Kawaguti 1932). On the one hand land planarians have practically no water-saving adaptations, while on the other hand they avoid wet environments and thus are considered to be stenohygric hygrocoeles (Froehlich 1955). It should here be noted that the situation may not be as absolute as suggested by Kawaguti’s (1932) findings, in that land flatworms may prevent desiccation by encasing themselves in mucus. Further, there may be also biochemical adaptations (secondary metabolism; see Campbell 1965) in land planarians for the elimination of carbon dioxide and nitrogenous waste that appear to facilitate water retention, such as the secretion of calcium salts (Percival 1925), which is consistent with being a stenohygric hygrocoele. In contrast to land planarians, freshwater species are subjected to osmotic influx of water and therefore must continuously regulate its volume with the help of their protonephridial system.

One way to assess the hydration of the planarian body is to determine the osmolarity of the tissues, which is expected to vary inversely with the degree of hydration (Jones et al. 2004). The osmolarity of the freshwater planarians Schmidtea polychroa and Girardia dorotocephala (Woodworth, 1897) was 125–128 mOsm kg⁻¹ and 126 mOsm kg⁻¹, respectively, while that of the marine triclad Procerodes littoralis (Ström, 1768) ranged between 217–272 mOsm kg⁻¹ (Jones et al. 2004 and references therein). Tissue osmolarity for the terrestrial planarian Arthurdendyus triangulatus (Dendy, 1896) ranged between 187.8–257.5 mOsm kg⁻¹, depending upon the laboratory conditions (relatively dry or fully hydrated) under which the specimens were kept (Jones et al. 2004). These data show that tissue osmolarity in land flatworms is higher than that of freshwater planarians. This means that terrestrial flatworms are less hydrated than freshwater species and that the former are able to maintain a higher osmolarity because they are less prone to influx of water. Evidently, the necessary amount of water that land triclads require for upholding their respiratory physiology is extracted from their frequently, but not always, humid environments. It is a matter of fact that for already millions of years land planarians have successfully evolved the necessary adaptations for exploiting the terrestrial environment but that our understanding of these solutions is still in its infancy.