June 23, 1893.] 



SCIENCE. 



339 



Tristan d'Acunha, and Juan Fernandez if it was not carried there 

 by man; yet it already shows some divergence from the type, and 

 the specimens from the two latter islands, though they are so far 

 apart, are alike. This is not extraordinary if we assume that 

 like climatic conditions produce like effects, since the two islands 

 are both far out in the ocean, at about the same parallel. 



The problem becomes complicated, however, when we find 

 Amalia gagates reappearing on the Pacific coast of North Amer- 

 ica, apparently quite native, though separated by long distances 

 from other localities for the species. This Pacific form generally 

 goes under the name hewstoni, given by Dr. Cooper, but I have 

 examined authentic examples, and am convinced it is only gagates. 

 Nor is this all, for in Australia and New Zealand are species of 

 Amalia so very near gagates that some recent students have 

 merged them in it. I have examined A. antipodanim, Gray, A. 

 emarginata. Hutton, and A. fuliginosa, Gould, from New Zealand. 

 A. emarginata I consider certainly a form of antipodaruni, but 

 this and fuliginosa appear to me to be valid species. Tbey very 

 much resemble gagates in structure, it is true, but, if they are 

 really the descendants of imported slugs, the amount of modifica- 

 tion they have undergone is remarkable. A. fuliginosa is in the 

 British Museum also from the ''Polynesian Islands" — exact 

 locality not stated. There is also an Amalia in the Sandwich 

 Islands, evidently very near to gagates. but whether identical with 

 it or an endemic form cannot be ascertained in the absence of 

 specimens. 



Thus it is seen that Amalia gagates and its allies present to us 

 some curious problems, which can only be solved by the collection 

 of specimens from many localities, and their very careful com- 

 parison. Because the slug was described from and abounds in 

 Europe, it does not therefore appear certain that specimens found 

 in distant localities, closely resembling gagates, are descended 

 from imported examples. We have often good reason for believ- 

 ing that this is their origin, but there is none of the certainty that 

 we feel in regard to other species now found at the antipodes. 

 Quite a similar example is afforded by Agriolimax loevis and its 

 allies, which seem certainly native in very widely-separated places. 

 It seems that A. gagates and A. loevis are very ancient species, 

 surviving in those places where the climate suits them. 



A STUDY IN POLARIZATION.— PRELIMINARY NOTE. 



BY JOHN DANIEL, VANDEREILT UNIVERSITY, NASHVILLE, TENN. 



Using a voltameter with platinum electrodes, separated by a glass 

 partition bored in the centre with a hole two centimetres in 

 diameter, over which was sealed a smaller glass plate bored with a 

 hole one and one-half centimetres in diameter, this smaller hole 

 being covered by metal plates of various thicknesses sealed tight 

 over it, a study has been made of the polarization phenomena 

 upon these thin metal partitions in different electrolytes and un- 

 der various conditions as to thickness of partition, current 

 strength, temperature, etc. 



Without now going into details of the apparatus, methods, and 

 results, the following summarized statement may be interfst- 

 ing: — 



1. The polarization on a gold-leaf partition in good-conducting 

 HjSO^ is zero, or too small to detect with our apparatus, for the 

 range of current used. 



3. The "critical thickness" in good -conducting solutions of 

 H^SO^, CuSOj, and NaCl is greater than .00009 millimetres for 

 gold; .00015 millimetres for platinum; and .0005 millimetres for 

 aluminum, under the above conditions. It is less than .0004 

 millimetres for gold; .003 millimetres for platinum; and .003 

 millimetres for silver. 



3. The " upper crkical limit" of thickness under these condi- 

 tions seems to be about . 004 millimetres, rather less than No. 3 gold. 



4. Tables I., II., and III. all point to the conclusion that be- 

 tween " critical limits" of thickness the polarization for a given 

 current increases with the thickness. 



5. Table II., showing relation of polarization to current, ex- 

 presses two interesting facts : (a) that the polarization on " thick " 

 plates is about the same, in this voltameter, for all currents be- 



tween .3 ampere and, say, .01 ampere, provided time enough be 

 allowed in each case for the current to become constant, i.e., be- 

 tween the upper limit of current, at which the development of 

 gas is so profuse as by mechanical obstruction and irregular 

 escape to interfere, and the lower limit, at which the formation 

 of gas is no faster than it can be dissipated. (6) Quite different 

 is the case for "thin" plates, where, within the limits of current 

 and thickness prescribed, the polarization is dependent upon the 

 current and gives for each thickness a different curve, or rather 

 straight line, for they are all straight lines converging to the 

 origin, and differing only in slope. The current strength at 

 which the polarization on very thin plates would reach a maxi- 

 mum is far above that used, being, perhaps, expressed in amperes 

 instead of tenths and hundredths. 



By thick plates are defined those above the "upper critical 

 limit; " by thin plates, those below this limit of thickness. 



6. Inspection of Table III., which gives the time-change of the 

 polarization, will show a similar distinction between "thick" 

 plates and "thin" plates, as was noted in the last paragraph, 

 viz., that for thick plates the change is considerable and continues 

 slowly for hours; for thin plates, the change of polarization with 

 time is both less pronounced and extends over much less time. 



7. It was noted, especially in the case of CUSO4 as electrolyte, 

 that there was polarization on gold-leaf if the gold exposed came 

 in contact with the solution some distance beyond the edge of the 

 hole in the glass plate to which it was sealed; thus in CuSO^, for 

 the stronger currents used, there was a symmetrical deposit of 

 Cu, decreasing in thickness from the outside toward the centre, 

 and vanishing at a small distance from the edge of the hole, this 

 distance being less the stronger the current. If only one corner 

 was left exposed, the Cu was deposited there This phenome- 

 non was further tested by bending a thick strip of aluminum, 4 

 centimetres long, into the shape of a narrow U, and simply hanging 

 this U in the ojsen hole of the glass partition, in CuSO.,, and closing 

 the circuit on the voltameter; the two ends of the metal strip 

 being thus in contact with the CuSO^ on opposite sides of the 

 glass two centimetres from the edge of the opening, there was 

 decided deposit of Cu on one end and escape of oxygen from the 

 other end. 



8. In CUSO4, all the plates except those below the critical 

 thickness were destroyed by oxidation. No. 1 silver was destroyed 

 is less than one minute. Of course, gold and silver above the 

 critical thickness could not be used in NaCl. because of chemical 

 action, though the thinnest plates were quite unaffected. Only 

 the No. 7 gold was tested in KOH, as it dissolved the sealing- 

 wax. 



9. Thick plates of gold were strongly oxidized in H2SO4, es- 

 pecially with strong currents. Thin gold plates were apparently 

 only oxidized under action of strong or long-continued currents. 

 Compare Tables II. and III. Silver was even more easily oxidized 

 than gold. Aluminum was so intensely oxidized by the current 

 that no satisfactory measurements could be made for this metal, 

 though the thin foil was unaffected. 



10. With H5SO4 as electrolyte, after a thick plate of pure gold 

 had been used as partition for the time-change of Table III., the 

 end cathode was found to be gilded. A thick Pt plate being 

 then substituted for the gold in the sanie solution for the results 

 of No. 1 Pt in Table III., the Pt partition was found, on re- 

 moval, to be gilded. The polarization for No. 1 Pt in this case 

 was somewhat less than for the same Pt after both it and the 

 end electrodes were thoroughly cleansed, the electrodes re-platin- 

 ized, and fresh solution made. 



11. The polarization in CuSOj, using Cu electrodes, reached 

 a maximum almost immediately and remained very constant. 

 The maximum polarization for thick Pt in CUSO4 was hardly 

 75 per cent of that for the same in HjSOj. In NaCl the polariza- 

 tion became constant very quickly also, but its value was de- 

 cidedly greater, especially on thin plates, than in HaSO^; 

 though the same distinctive behavior of thick and thin plates 

 was maintained. 



12. In H2SO4 of different concentrations the maximum polari- 

 zation for a partition was of the same order of magnitude; but its 

 value for very weak currents was decidedly greater in weak solu- 



