segments contain sexual products. In the male these sexual segments are generally pink, while in the female they are dull sagegreen or brownish yellow. These colors are entirely due to the color of the contained sexual elements, however, for when the eggs or sperm are discharged the swimming segments of both sexes are dull brick-red in color.

When the swarm occurs the hinder end of the worm crawls out backwards from the burrow, and attempts to swim away from the anterior non-sexual part which remains within the burrow and takes no part in the swarm. A constriction appears allowing the sexual portion to break away from the anterior part of the worm; and immediately the posterior end swims with great rapidity almost vertically upward to the surface, upon reaching which it continues to swim hind end foremost parallel with and very near to the surface of the ocean. This occurs at least two hours before sunrise for I found the Palolo swimming at 3.20 A. M. on July 25, 1902. The worms swim in all directions and begin to discharge sperm or ova through their nephropores, bu^ the least stimulus, such as being lifted from the water, or the current from the stroke of an oar, will cause them to contract violently, often breaking themselves into fragments and casting the sexual products out through rents in the cuticula. This normally occurs as soon as the first rays of the sun fall upon the water, and in a few minutes after sunrise all of the worms will have completely freed themselves of genital products, and the ocean becomes milky in appearance due to the vast quantity of sperm and ova floating within it. The worms continue to swim but they gradually sink down and disappear so that none are to be seen two or three hours after sunrise. In a dense swarm there may be on an average about one worm per square foot over wide areas of the sea. The worms are commonest in places where the water is about 6 fathoms in depth. Great numbers of the worms are devoured by fishes as they sink, although they arc not attacked to any great extent while upon the surface. When set free the sexual ends swim vertically upwards with such rapidity that they run little risk of capture, and this habit must be a great advantage to the worm. If the swimming worm' be broken into fragments each piece continues to swim backward in a normal manner, showing that the reaction is not controlled by any one ganglion or localized group of ganglia, but that the whole sexual end of the worm is affected by the stimulus which causes the breeding-swarm.

The females are iruich more sensitive than the males. Males may readily be killed in weak alcohol or formalin without casting out their sexual products, whereas it is difficult to preserve perfect females in this manner. A count of 180 worms collected at random gave 103 males and 77 females. It appears, therefore, that there are about 57% of males and 43% of females.

Growth and Development.

The eggs float at or near the surface, and are in the 4-16 cell stages at about 7 A. M. The segmentation is total and unequal, and the gastrula is formed by epibole. The larva is telotrochal, and possesses two eyes and large ectodermal cephalic glands. It swims through the water until four pairs of setigerous appendages have developed, after which it sinks to the bottom. The development was described in detail by the author in 1900.

During June and July, 1902, large numbers of coral rocks were broken open and many hundreds of the complete worms were obtained. The worms appeared to be of two sizes, small slender ones about 90-200 mm. long, and large stout ones 250-350 mm. in length.

The great majority were slender and about 200 mm. in length and showed no signs of being sexually mature, while the large worms were all developing sexual products in their posterior segments. Only about one worm in a dozen was developing sexual products, and on July 28 after the last swarm I broke open a large number of rocks and obtained 43 worms. Three of these were of large size and their posterior ends were constricted, the sexual segments having evidently been recently cast off. The constricted ends had healed but no regeneration had taken place. (See figure 7.) The remaining 40 worms were 140-225 mm. in length and 36 of them were complete, two were regenerating at the posterior end of the body, and two had recently lost the extreme posterior segments apparently through accident in being cracked out of the rock, for the ends had not yet healed. All of these 40 worms were shorter and more slender than those which had evidently possessed sexual segments, and none of them were sexually mature. It appears, therefore, that the worm requires at least two years to attain sexual maturity, and that only about one worm in a dozen gives rise to sexual segments in any one year. This may represent the ratio of survival of the worms, only about one in twelve living to be two or more years old.

It seems reasonable to suppose that the anterior part of the worm regenerates a new posterior end after the swarm. Many regenerating worms of all sizes are found, and, indeed, I had two worms whose posterior ends were broken off on June 25, and maintained them alive within coral rocks kept beneath low tide level until July 28, and during this period they had regenerated slender, tapering posterior ends about 4.5 mm. long (see Figure 6).

Experiments on Fertilization, Etc.

When rocks containing the Atlantic Palolo are cracked open the worm begins to crawl out backward, but when about half of the length of the worm has emerged, the anterior part attempts to return into the burrow while the posterior part struggles to swim backward. A constriction appears between the two ends, and the posterior part of the worm is set free and swims through the water, while the anterior crawls back into the rock. In large worms with well developed sexual segments the break is apt to appear exactly at the junction between the sexual segments and the non-sexual anterior part of the worm. The sexual segments then swim rapidly backward and upward to the surface, where they continue to swim backward with great activity. Soon, however, they begin to contract violently, casting sperm or ova 6ut into the water very much as in a normal swarm. It appears that the shock given to the worm upon cracking the rock acts as an abnormal stimulus to produce the drama of the breeding-swarm, which can be produced in this manner at any time of the day. On July 12, 17, 19 and 23, 1902, I obtained eggs and sperm from worms which had thus been induced to go through with the actions of a swarm, but although the sperm was very active, and appeared to be normally attracted by the eggs, not one living embryo was obtained, although many of the eggs segmented abnormally. It was remarkable that eggs obtained at 4 P.M. on July 23 could not be fertilized although less than 12 hours afterwards the normal swarm occurred in which practically every egg was fertilized. This is clue to the fact that none of the eggs become mature until the time of the normal swarm when all of the eggs mature simultaneously. Full-grown female worms which had discharged some eggs were killed on July 12 and 23, at 4.30 P.M., in a saturated solution of corrosive sublimate in 35% alcohol; and four females which were swimming in the swarm at 4.20 A.M. on July 24 were similarly treated. Figure A shows the condition of the eggs in the females killed on July 12 and 23. The nucleus is central in position, the nuclear membrane is complete, and there is a large nucleolus. Figure B, on the other hand, shows the condition of the eggs within the females killed during the swarm at 4.20 A.M. on July 24. The eggs are now about to give off the first polar globule. The nucleus has migrated to the surface at the animal pole, the nuclear membrane has disappeared, the nucleolus has broken up into small spherical fragments,* and the first polar globule is not yet constricted off from the egg.

[graphic][ocr errors]

I hoped to discover whether the Atlantic Palolo responds in its ■ swarm to the light, or to some attraction due solely to the position of the moon. Accordingly a number of worms were maintained alive within coral rocks beneath low tide level, and on the nights of July 23-28, 1902, the rocks containing the worms were placed in glass aquaria which were shut up within a wooden boat shed into which no direct moonlight could penetrate. None of these worms swarmed, but upon cracking open the rocks I discovered that all of the worms were immature. The experiment is therefore inconclusive and should be repeated.

In 1900 I advanced the view that the Palolo may be regarded as an animal in which the egg-laying season has been reduced to a period of one or two days; and that it could be shown mathematically that a shortening of the egg-laying period causes a greater concentration of breeding individuals and therefore shortens the average distance that the spermatozoa must travel in order to fertilize the

*Xot shown in this individual section.

ova. This advantage would allow of an increase in the relative number of females and a reduction in the number of males, and thus more eggs would be produced. The short breeding season of the Palolo may, therefore, have been brought about through the agency of natural selection, provided the ancestors of the Palolos now living possessed an initial tendency to swarm at or near the period of the moon's last quarter rather than at other times, and that this tendency was of a sort that would be perpetuated by inheritance.

Friedlaender, 1901, combats this view, holding that the Darwinian explanation is inadequate. He maintains that the Palolo responds to a certain condition of the tide in conjunction with an influence which the moon exerts upon the worm when at the position of last quarter, and that the response of the worm has not been brought about through natural selection.

It is interesting to observe, however, that the condition of the tide at the Tortugas, Florida, upon the day of the last quarter of the July moon is quite different from its condition at Samoa and Fiji upon the days of the last quarter of the October and November moons. At these times in Fiji and Samoa the tide is about at its lowest when the sun rises, while at the Dry Tortugas the tide is ebbing but is still about l*4"2?4 hours high. For example: On July 27, 1902, at the Dry Tortugas low tide occurred 1 hour and 22 minutes after sunrise, while at Apia, Samoa, on October 23 and November 21, 1902, low tide came at 1 and at 14 minutes after sunrise on the respective dates. Also the mean rise and fall of the tide at the Dry Tortugas is only 1.1 feet, while at Samoa and Fiji it is 2.2—3.7 feet. All that we can say is that Atlantic and Pacific Palolos swarm upon or near the day of the last quarter of the moon, at corresponding seasons of the year, for the meteorological conditions and temperature of the water at the Tortugas in July are remarkably similar to those at Samoa and Fiji in October and November. The worm no doubt responds to some physical stimulus which is dependent upon the condition or position of the moon, but the exact nature of this stimulus remains to be discovered.

Deseription of the Mature Worm.

A minute description of Eunice fncata is unnecessary, as the worm has been well described by Ehlers, 1887; Mem. Mus. Comp. Zool. at Harvard Coll., Vol. XV, p. 91, PI. 25. Ehlers, however, had

« PreviousContinue »