The sequence of origin of the postmetamorphic rays in Heliaster and
Labidiaster (Echinodermata: Asteroidea)

Origen sequencial de los rayos postmetamorficos en Heliaster y Labidiaster
(Echinodermata: Asteroidea)

PATRICIO SANCHEZ R*

ABSTRACT

Postmetamorphic rays in Heliaster and Labidiaster originate in four `quadrants' between the five primary rays, and not normally in the madreporic interradius. The rays originate in one of two very definite sequences depending on the species. H. kubiniji and H. multiradiatus share one sequence, whereas H. canopus, H. helianthus and Labidiaster share the other sequence. Pycnopodia, Rathbunaster, Heliaster and Labidiaster add rays at far greater sizes and in a manner that is distinctive from other multiradiate starfish, possibly indicating a new taxonomic unit.

Key words: Echinodermata, development, symmetry, Heliaster, Labidiaster.

RESUMEN

Los rayos post-metamórficos en Heliaster y Labidiaster se originan en cuatro"cuadrantes" entre los cinco rayos primarios, y no como ocurre normalmente en el interradio del madreporito. Los rayos se originan en una de dos secuencias definidas, dependiendo de la especie. H. kubiniji y H. multiradiatus comparten una secuencia, mientras que H. canopus, H. helianthus y Labidiaster comparten otra secuancia. Pycnopodia, Rathbunaster, Heliaster y Labidiaster adicional rayos de tamaños mucho mayores y de una manera que es distintiva de aquel de otras estrellas de mar multiradiadas, posiblemente indicando su pertenencia a una nueva unidad taxonómica.

Palabras clave: Echinodermata, desarrollo, simetría, Heliaster, Labidiaster.

Santiago, March 9, 1976

I did in fact investigate the origin of new, postmetamorphic rays in Heliaster during a residence at the MCZ [Museum of Comparative Zoology, Harvard University], back in 1958-59. I was led to this on finding a specimen of H. helianthus, in Chile, with several small rays regularly spaced every third large ray. Reviewing the literature, HL Clark's (1907) conclusion that rays originate with no definite order seemed untenable, in view of my specimen. Yet it took me a long time and the examination of very many specimens before I came to the answer, for the sequence is not simple. I reported my findings in seminars and lectures at several universities in the USA at that time (see: Emery Swan's comment on page 410 of Boolootian's Physiology of Echinodermata) (Swan 1966, Boolootian 1966) and I presented a formal account at the 1st. Chilean Zoological Congress in Santiago, September of 1960. As things are in this part of the world, the Proceedings of this Congress were never published completely and my report has remained unpublished so far. I myself became involved in other projects, which prevented me from preparing a text to be published elsewhere. Moreover, I was hopeful at that time to make a sizable collection of really young specimens in order to settle one final question which had remained unanswered, in spite of the collections I had examined in the USA - including one I left at the MCZ - which were mostly of large or medium size Heliaster. It turned out that very small specimens are difficult to find in large numbers and it is only after years of field work that I have of late come across a `nursery ground' which has provided the material necessary for the purpose. So I am now back to this research and giving the final touches to my manuscript.

The sequence of origin of postmetamorphic rays is not simple in Heliaster, being `obscured' by at least two factors: individual variability and regeneration of rays. Yet I will try to provide you with the core of the answer, in anticipation of the full story.

1. HL Clark (1907) was correct in assuming that new rays originate in four of the original interradial spaces (A-B, B-C, D-E, E-A) and not normally between the original rays C and D, the five original rays being those with gastric muscles (rays are lettered such that the madreporite is between rays C and D). He was also correct that certain territories or quadrants have different activity rates, in that new rays appear earlier and/or grow faster in some of the original interradial spaces, as illustrated in his plate no. 8. It is my conviction that Clark was mistaken in concluding that rays originate in no definite order in those four interradial spaces. Furthermore, it seems to me that the `gradients' of new-ray-growth found by Clark in H. kubiniji (his plate 8) do not hold alike for all the species of the genus.

2. New postmetamorphic rays originate in very definite sequences in the original interradial spaces A-B, B-C, D-E and E-A but, surprisingly enough, the sequence is not exactly the same in all Heliaster species. In H. kubiniji and H. multiradiatus, the sequence is the same as described by Ritter & Crocker (1900) for Pycnopodia helianthoides, but in these Heliaster species the process occurs in four regions, as mentioned. In all the other species (including H. canopus, which grows up to some 25 rays, as do H. kubinji and H multiradiatus), the sequence is another. I trust the difference can be easily grasped on examining the illustration from my manuscript which is enclosed (Fig. 1). As this illustration shows, the `age' of the rays can be recognized by their differences in length (very small rays not shown there) and also by the alignment of the marginal aboral spines in the proximal portion of rays, even if of equal size. New rays are numbered 1, 2, 3 in this figure, implying the order of their appearance in each original interradial space (B-C in this figure); first one grows, then pairs, but this last statement requires qualification (see below).

3. In H. kubiniji, H. multirariatus and H. canopus, three `generations' of new rays growing in four of the quadrants bring the total number of rays to 25, which is the `normal' for these species. Yet further rays can grow and they normally appear according to the pattern of each of the species; that is, in the angle between the youngest pre-existing rays and the originals (A,B,C..) in H. kubiniji and H. multiradiatus, whereas they appear between the two youngest rays of each quadrant in H. canopus. More than three `generations' of new rays are needed to complete the process in all the other species, and they grow following the pattern described for H. canopus. A fourth generation of 8 new rays brings the total number of rays to 33; a fifth generation, to 41. Yet these `final' numbers also require qualification.

4. Several qualifications are required in order to fully describe what occurs in each individual specimen. The most important is this: the schemes or patterns previously described indicate the relative position of subsequent `generations' of new rays in each of the four interradial spaces A-B, B-C, D-E and E-A, for the two groups of species, but they do not imply that the members of each generation are strictly synchronous in their appearance, neither within nor between quadrants. In fact, there are cases where all rays of a given generation are of about equal size, smaller than the rest; these cases are rare but such was the specimen mentioned above which put me on this problem. In most cases, one member of a pair or generation is notably smaller, or even absent, in a quadrant and the smaller - or absent - member may be any of the pair. As recalled, Clark found that normally the process of ray multiplication is more advanced in some quadrants than in others and as mentioned, I have evidence indicating that these `gradients' differ in both groups of Heliaster species. At any rate, what occurs in any one specimen is that one quadrant may, for example, be producing rays of 2nd. generation while another quadrant may have some of 3rd. or 4th. As a consequence of this variability, it is infrequent to find specimens, even with the same total number of rays, that are exactly alike in the position of their new rays and in the relative length of these. I have no indications as to differences in growth rates in different quadrants nor in different generations of new rays, which would certainly complicate things. Assuming no such difference is the simpler hypothesis but, even so, the variability observed in new ray appearance would seem to contradict my conviction of a definite sequence in specimens as a whole (beyond the sequence within each quadrant, illustrated in my figure). This conviction stems from my previous observations, but could not be proven with the samples then available, large as they were. I trust that statistics of large samples of small specimens, as now available, will enable a solution to the question. This should also settle the question of differences in growth gradients in Heliaster species-groups.

5. Regeneration of rays is a frequent event in Heliaster. It can involve larger or smaller fragments of isolated rays, or of two or three adjacent rays, or it can involve the whole length - from their origin in the disc proper, on the aboral side - of a series of adjacent rays as shown in the photograph enclosed (Fig. 2). This last feature is in the literature but it is only recently known that it is due to autotomy in Heliaster. Dr. Carlos Antonio Viviani, a young colleague and former associate of us has found that H. helianthus reacts thus to the predation by the voracious Chilean sea star Meyenaster gelatinosus; a clever adaptation indeed! (see Viviani 1978). At any rate, with experience, regenerating rays can be distinguished from truly new ones, so this is no major confusing factor in tracing back the history of origin of rays in a specimen with regeneration. But what can be important is that regeneration seems to inhibit the process of new ray appearance, at least in its vicinity. Yet this can only be proven by experiments, which I expect to do now that we have sea-water aquaria in our laboratory.

6. It is known that different species of Heliaster attain different numbers of rays, these being centered around ± 25 rays (H. kubiniji, H. multiradiatus and H. canopus) and ± 35 rays (all other species), yet with high intraspecific variability. It is also known that the origin of new rays diminishes in frequency with increasing size. There is a correlation between size and number of rays, which is different in the different species, and usually with very high dispersal (my data). The result is that beyond a given size, no further new rays arise, the individuals remaining with as many as could grow `in time'. How much of this is due to genetic variability (for `number of rays'), or to growth rates (genetically or environmentally controlled), I do not know. But it does seem possible to me that an inhibitory effect of regeneration, which is common in Heliaster, could partly account for the observed individual, intraspecific variation in the final number of rays in these species.

7. May I finally add that the resemblances between the patterns of new ray growth in Pycnopodia helianthoides and some Heliaster species seem to me of some significance, for I have good evidence that other multiradiate forcipulate sea stars (Labidiaster, Rathbunaster), also increase in number of rays by the same patterns as found in Heliaster. But that is another story.

*************************************

Santiago, June 26, 1976

I have been in Uruguay, Brasil and Argentina for the last three months. I have returned recently to Santiago but only for a few days, on my way to Panama and México, from where I shall return in mid-August and remain here, to finish the Heliaster study. So I shall now comment only on some points.

My use of letters C and D for the rays between which occurs the madreporic plate follows Hyman's (1955) explanation of Carpenter's system. I was not aware of Moore & Fell's (1966) discussion and am grateful to learn about it. I have not checked on the position of the anus in Heliaster, and it seems, from my recollection, that it would be difficult to relate it to interradial spaces for it is small and central on the disk, but I shall try to determine this.

The news that Acanthaster also produces post-metamorphic rays is also new to me; I was unaware of Yamaguchi's (1973) research and my examination of figures of Acanthaster in the literature had convinced me that small rays in this genus were due to regeneration. The news is somewhat of a disappointment, for I was in the belief that only some isolated families of Forcipulata showed this feature (Pycnopodia, Rathbunaster, Heliaster and Labidiaster), which could have made a good case for a new taxonomic unit. Acanthaster lacks the first, unpaired ray which is present in H. canopus in each of the four quadrants that produce new rays (number 1 in Fig. 1). Furthermore, the next new rays in sequence, which are in pairs (2-2, 3-3, in Fig. 1), do not bear the same positional relations in Acanthaster and H. canopus.

The following diagram Figs. 3 & 4, continues the process beyond my drawing Fig. 1 of H. canopus. Such continuation occurs in H. helianthus and in other species with high numbers of rays but of the same pattern. It seems to me that this H. canopus pattern does not occur in any of the genera that have been previously studied, but I am glad to be able to confirm now that Labidiaster also shares this same pattern. In fact, during my recent visit to Argentina I was able to examine a satisfactory collection of L. annulatus from Antarctica, which provided the evidence for the previous statement.

As to the evolution of the different patterns, I am unable to choose between alternative hypotheses. A switch from one to another pattern in species that are morphologically and ecologically so similar as the different Heliaster species seems to me as unlikely as their independent origin. We would have to know more about the genetic and developmental basis of these events. The conjecture that polymerism first developed in one interradius and then extended to others, makes sense it seems to me, but what excludes the possibility that the process could have been generalized to start with, and secondarily restricted to the anal interradius? Further knowledge on gradients between quadrants may contribute to the question. Also, of course, knowledge on the case in fossil multiradiate asteroids.

Editor's explanation and postscript

The text of this posthumous paper was transcribed with minor editing from letters written by Prof. Sanchez to FHC Hotchkiss. Copies of the original correspondence are in the possession of JC Castilla and FHC Hotchkiss. The species mentioned are attributed to the following authors (see Madsen 1956, Maluf 1988): Heliaster kubiniji Xantus 1860, H. multiradiatus (Gray 1840), H. canopus Perrier 1875, H. helianthus (Lamarck 1816), H. cumingii (Gray 1840), Pycnopodia helianthoides (Brandt 1835), Meyenaster gelatinosus (Meyen 1834), Labidiaster annulus Sladen 1889. Prof. Sánchez was correct that undersized rays in Acanthaster are due to regeneration in specimens larger than approximately 20mm overall diameter. In Acanthaster the supernumerary rays form only while it is at a very small size, attaining 16 to 18 rays when only 12mm overall diameter. It was an insight of Prof. Sanchez to regard it significant that other multiradiate forcipulate starfish (Labidiaster, Rathbunaster and Pycnopodia) increase in number of rays by the same patterns as found in Heliaster. This argument for a new taxonomic unit is still intact today. The discovery by Prof. Sánchez that Labidiaster follows the same pattern as Heliaster and adds posmetamorphic rays in just four quadrants gives a probable explanation for the color pattern of Labidiaster that is seen in undersea photographs published by Dearborn et al. (1991). The color pattern has four distinctive dark quadrants that separate what must be deduced to be the five primary rays, with a primary ray on each side of the madreporite. AH Clark (1939) described an H. cumingii with new ray buds that tunnel from the oral to the aboral surface of the disk to take their places between preexisting rays. Although Heliaster is distributed at present only on parts of the eastern Pacific shoreline, fossils of Heliaster were found in the Pliocene of Florida (Jones & Portell 1988). The relationship between number of rays and body size diameter in H. helianthus based on a large sample was graphed by by Tokeshi et al. (1989: Fig. 2). Hotchkiss (in press) undertakes a new analysis of the developmental basis that underlies the number of rays in starfish; that paper is dedicated to the memory of Prof. Sánchez. FHC Hotchkiss. Thanks Rich Mooi, Mark Reilly, JC Castilla, John Lawrence, John Dearborn, Cecelia McIsaac and Anita Hotchkiss for assistance, literature and data.

* Passed away on February 23, 1999. This paper was edited by Frederick H. C. Hotchkiss, 26 Sherry Road, Harvard, MA 01451 USA, from two letters.

LITERATURE CITED

BOOLOOTIAN RA, ed (1966) Physiology of Echinodermata. Interscience Publishers, New York. xviii + 822 pp.

CLARK AH (1939) Echinoderms (other than holothurians) collected on the Presidential cruise of 1938. Smithsonian Miscellaneous Collections 98(11): 1-18, 1-5 pls.

CLARK HL (1907) The starfishes of the genus Heliaster. Bulletin of the Museum of Comparative Zoölogy at Harvard College 51:25-76, 8 pls.

DEARBORN JH, KC EDWARDS & DB FRATT (1991) Diet, feeding behavior, and surface morphology of the multi-armed Antarctic sea star Labidiaster annulus (Echinodermata: Asteroidea). Marine Ecology. Progress Series 77:65-84.

HOTCHKISS FHC (in press) On the number of rays in starfish. American Zoologist.

HYMAN LH (1955) The invertebrates, Vol. IV. Echinodermata. McGraw-Hill, New York.

JONES DS & RW PORTELL (1988) Occurrence and biogeographic significance of Heliaster (Echinodermata: Asteroidea) from the Pliocene of Southwest Florida. Journal of Paleontology 62: 126 - 132.

MADSEN FJ (1956) Reports of the Lund University Chile Expedition 1948-49. No. 24. Asteroidea, with a survey of the Asteroidea of the Chilean shelf. Lund Universitets Årsskrift, N.F. Avd. 2, Bd. 52, Nr. 2. (Kunglig Fysiografiska Sällskapets Handlingar, N.F. Bd. 67, Nr. 2) 53 pp. + 6 plates.

MALUF LY (1988) Composition and distribution of the Central Eastern Pacific echinoderms. Natural History Museum of Los Angeles County Technical Reports No. 2, 242 pp.

MOORE RC & HB FELL (1966) Homology of echinozoan rays. In: Durham JW et al. Echinodermata 3, Asterozoa and Echinozoa. RC Moore (ed) Part U of Treatise on invertebrate paleontology: U119-U131. Geological Society of America and University of Kansas Press, New York.

RITTER WE & GR CROCKER (1900) Multiplication of rays and bilateral symmetry in the 20-rayed star-fish, Pycnopodia helianthoides (Stimpson). Proceedings of the Washington Academy of Sciences 2: 247 - 274.

SWAN EF (1966) Growth, autotomy, and regeneration. In: Boolootian RA (ed) Physiology of Echinodermata: 397-434. Interscience Publishers, New York.

TOKESHI M, C ESTRELLA & C PAREDES (1989) Feeding ecology of a size-structured predator population, the South American sun-star Heliaster helianthus. Marine Biology 100:495-505.

VIVIANI CA (1978) Predación interespecífica, canibalismo y autotomía como mecanismo de escape en las especies de Asteroidea (Echinodermata) en el litoral del Desierto del Norte Grande de Chile. Published by the author, Iquique, Chile. 116pp + 25 plates.

YAMAGUCHI M (1973) Early life histories of coral reef asteroids, with special reference to Acanthaster planci (L.). In: Jones OA & R Endean (eds) Biology and geology of coral reefs, vol. 2, Biology 1: 369-387. Academic Press, NY.

Invited Editor J.C. Castilla