634 
NATURE 
[AuGUST 22, 1912 
Station, Dr. C. G. Joh. Petersen. How necessary 
both are can be judged from the sequel. 
In 1905 we continued our work in the Atlantic 
between Iceland and Brittany. The larve of the 
eel were found in quantities, as many as seventy 
in a single haul, from the Hebrides southwards, 
but not to the north and not to the east of the 
1000 m. line—thus not in the North Sea or Nor- 
wegian Sea. I drew the conclusion from this 
(1906), that all the eels which occur in the North 
[European countries must come from the Atlantic. 
A comparison of all the available data for the time 
of appearance of tlie elvers in the rivers of West 
and North Europe confirmed me in this conclusion. 
In 1906, May-June and August-September were 
devoted to an investigation of the waters between 
North Spain and South-West Ireland. No 
younger stages than those of 1905 were found, 
but it proved that the larvee were not restricted 
to the belt between 1000 and 2000 m., where we 
had found them in 1905. They were taken out 
over 5000 m., the greatest depths investigated. 
More than 500 specimens were taken, and a curi- 
ous thing was that the spring specimens were not 
yet metamorphosed, whilst those taken in autumn 
were for the most part in process of transforma- 
tion. It proved, further, that the youngest speci- 
mens (not metamorphosed) occurred furthest out 
to sea, the oldest metamorphosing nearer the 
coasts (see Fig. 1). 
It was thus perfectly clear that the larve of 
the eel are not demersal fishes, as Grassi and 
Calandruccio had supposed, but belong to the 
surface waters, even out over the greatest depths. 
Nevertheless, we had not yet found the early 
stages, and I was now (1909) inclined to the belief 
that the larvae must be hatched out over great 
depths, far from the coasts and away from the 
bottom of the sea. This may be said to mark 
the end of the second stage in our investigation 
of the eel problem. 
A study of the distribution of the adult eel in 
the countries bordering on the Atlantic afforded 
an instructive commentary on our deep-sea investi- 
gations. As is shown in Fig. 2, the eel is quite 
wanting along the coasts of the South Atlantic, 
and does not even reach down to the equator, 
whereas it occurs on all the islands of the tem- 
perate North Atlantic. From an investigation of 
a large number of specimens, I found that the 
European eel (A. vulgaris) could always be dis- 
tinguished fromthe American eel (A. rostrata), 
from the fact that it has on an average seven 
more vertebra than the latter. The eels living on 
the Azores, for example, were found to be typical 
A. vulgaris. The explanation of this peculiar 
occurrence seemed to lie in the distribution of 
temperatures and salinities in the deeper layers 
of the Atlantic. In the Northern, temperate 
Atlantic and in the Mediterranean, thus, where the 
larve occur, these have higher values than in the 
Southern Atlantic. On the supposition, therefore, 
that the eel requires a high temperature and 
salinity for spawning purposes, we can explain 
the absence of eels on the coasts of the South 
Atlantic; further, the direction of the currents 
NO. 2234, VOL. 89] 
| there is such that the larve cannot be carried to 
the South Atlantic coasts (Fig. 2). 
The third stage in the solution of the eel 
problem began with our 1908 investigations in 
the Mediterranean and adjacent parts of the 
Atlantic. These investigations extended over both 
winter (1908-09) and summer (1910), so that all 
seasons of the year have been considered. The 
result has been a great surprise to me. Instead 
of finding here the youngest developmental stages 
of the eel—and Grassi’s publication in the Royal 
Society’s Proceedings of 1896 was certainly 
responsible for the general belief that the spawn- 
ing places would be found there—I have been 
obliged to come to the conclusion that the eel 
does not spawn in the Mediterranean at all. The 
Mediterranean owes its stock of eels, like the 
North Sea and Baltic, to the Atlantic. On the 
other hand, I have found the early developmental 
stages (eggs or early larvee) of fourteen other 
species of eels in the Mediterranean, which means 
that they must spawn there. This is in so far a 
distinct advance, for the eggs and early larve of 
| these forms could not previously be identified 
with certainty in any single case. 
With regard to the common eel, it is naturally 
more difficult to prove a negative result, e.g. that 
it does not spawn in the Mediterranean, than a 
pesitive one, e.g. that Muraena helena spawns 
there. It has been necessary, therefore, to take 
a number of conditions into consideration, of 
which the following are the most important. 
(1) Large quantities of eel larvae are carried by 
currents into the Mediterranean from the Atlantic. 
This has been proved by investigations on both 
| sides of the Straits of Gibraltar, and I have been 
able to follow their drift further east. The hydro- 
graphical investigations of Dr. J. N. Nielsen from 
| the Thor show that the surface waters, in which 
the eel larve live, are moving eastwards from the 
Straits of Gibraltar the whole year round. The 
current bottles I have had put out in the neigh- 
bourhood of the Straits further show that the rate 
of flow is at least twelve to eighteen miles in the 
twenty-four hours, as far east as ca. 11° E.L. at 
any rate. These observations prove incontestably 
that a portion at least of the eel stock of the Medi- 
terranean must come from the Atlantic. 
The same lines of reasoning, by which I showed, 
in 1906, that the eels of the North Sea and Baltic 
must come from the Atlantic—and the reasoning 
and facts have not been controverted—can also 
be applied to the Mediterranean. The larve of 
the eel were found by the Thor over the whole 
of the western basin (west of Italy), but increasing 
greatly in numbers towards Gibraltar, where the 
quantities were greatest. In the eastern basin 
(east of Italy) no larve of the eel were found— 
compare Italy with the 1000 m. line west of South 
Ireland. ‘The eel larve taken nearest Gibraltar 
were on an average smaller than those found 
further east, both by the Thor and in the collections 
I have had made for me at Messina during four- 
teen months (see Table). Further, at Messina, 
most larve were taken in spring and summer, 
fewest in winter, and this agrees with the fact 
