consin; nor that the Talbot represents about the same period of time. Both of these formations are singularly destitute of vertebrate fossils. On the other hand, the lowest terrace in Florida, Georgia, and the Carolinas is filled with remains of extinct vertebrates down to salt water. At Wilmington, N. C., the great sloth Megatherium and horses are found. The latter occur all along the coast of North Carolina, along the Potomac, and on the west shore of Chesapeake bay. The line of horsebearing localities is then taken up at Swedesboro, N. J., is continued past Philadelphia, and ends at the Navesink Hills. From the Potomac to Raritan bay it keeps far away from the Atlantic coast. In the Fish House clays, opposite Philadelphia, considerable horse remains have been found. By the New Jersey geologist these clays are regarded as belonging to the Pensauken formation; and this is referred to the early Pleistocene. The vertebrate fossils appear, therefore, to connect the lowest terrace of the south Atlantic states with the Pensauken, rather than with the Wisconsin. Berry's admission that the Vero deposits may be as old as the Peorian shows that he does not believe that any connection with the Wisconsin drift has been established.

The writer contends likewise that the Pensacola terrace has not yet been geologically correlated in the Mississippi Valley with any definite glacial stage.

Inasmuch as Berry grants that the Pensacola terrace may be as old as the Peorian interglacial stage he and I need have no quarrel about the age of the Vero muck bed. He may perhaps yet come to acknowledge that it may be as old as the Sangamon.

As regards Dr. Chamberlin's paper it may be stated that he has decided to abandon his theory of the secondary inclusion of the vertebrate fossils "unless all other explanations fail." He asserts (p. 667) that the dates of the appearance of man and of the disappearance of the extinct animals were among the very points brought into question and could not themselves be used as decisive criteria. With that part of this statement which concerns man I agree; but with that which re

gards the vertebrates I dissent. The time when those vertebrates lived and when they disappeared is to be determined by their relation to the deposits in which they have been found in a thousand or more other places in our country; and it is legitimate to apply the knowledge gained therefrom to the situation at Vero. Chamberlin seems to respect rather lightly the vertebrate fossils, for he believes that the time relations of the deposits were quite well indicated by the physical criteria, irrespective of their fossil contents. He believes, with Berry, that the marine marl bed and with it the Pensacola terrace is late Pleistocene in age. The writer takes this occasion to say that if the geologists can prove that proposition it will at once end the dispute about the time of the disappearance of the fauna represented at Vero; and vertebrate paleontology will become once more indebted to geology. Pending that proof I shall maintain, on the evidence of the vertebrate fossils, that that terrace belongs to the early third of the Pleistocene.

Dr. Chamberlin's faith in the value of fossils seems to be somewhat livelier when, in order to determine the age of the human relics at Vero, he cites the age of European pottery and men's bones; but what connection has been established between the use of pottery in Europe and its use in America?

It is not a little amusing to observe that the camels and horses and their fellows, which under the designation of a Pliocene fauna " were used at Table Mountain to combat the existence of early man, are now, at the other, far distant, end of the line, mustered in as a "mid-Recent fauna" and called into service to continue the same war.

OLIVER P. HAY CARNEGIE INSTITUTION OF WASHINGTON

SPECIAL ARTICLES

THE ANIMAL CENSUS OF TWO CITY LOTS

ASIDE from articles by McAtee, Banks and Herbert Osborn, very little attention has 1 McAtee, W. L., "Census of four square feet," SCIENCE, Vol. 26, pp. 447-49, 1907; Banks, N., "A 'census of four square feet,' '"' SCIENCE, Vol.

been given to the total animal population of definitely measured areas of land. At the suggestion, and with the aid and assistance, of Dr. V. E. Shelford, of the University of Illinois, a study was made of the animal population of two vacant lots, a block apart, in Champaign, Illinois, over a period from September, 1916, to May, 1917. The lots were on John and Green Streets, adjoining the Illinois Central Railroad. They had not been burned over for at least a year, and despite the encroachments of débris from the railroad and nearby houses, were essentially wild, being covered with grass, sweet clover, dandelions, burdocks and other weeds, while the Green Street lot had a thicket of young honey locusts on one side, and that on John Street had a brook bordered by osage orange and willows running along one side.

The apparatus used was very simple, consisting of a tin pail with the handle and bail removed, and a screw top, an inch in diameter, soldered into the bottom, through which the anesthetizing agent could be introduced. The sharp edges of the pail, where the bail 26, p. 637, 1907; Osborn, H., "Leaf-hoppers of Maine," Maine Agr. Expt. Sta. Bull., No. 238, pp. 81-160, 1915.

+ Indicates partly eggs.

had been cut away, would sink into soft earth, and with the aid of a big knife, vegetation and hard earth could be cut through, so that the animal population of that small area would be imprisoned in the inverted pail. A considerable amount of chloroform or ether was added through the screw top, and after clearing away the surrounding vegetation and débris, sufficient time would have elapsed so that all the active animals were anesthetized and the pail could be taken up without fear that any of them would escape. The vegetation within the circle was picked to pieces and shaken over newspapers, weed stems split up, the surface of the earth left bare was carefully examined and finally all the earth to a depth of six inches was dug up and sifted over newspapers, so that all animals large enough to be visible to the naked eye would be sorted out. The surface area covered by the pail was .518 sq. ft., which, multiplied by 84,092, gives the population per acre. Two examinations (one for each lot) were made daily when time was available, or the ground not too wet and sticky or frozen. The results are given in the table.

The area enclosed by the pail (.518 sq. ft.) is so small that there will necessarily be a wide

variation in abundance in all orders of animals in any two examinations, and it is only by averaging results for twenty-day periods that some idea can be gained of seasonal variation.

The great unevenness in the Orthoptera and Homoptera columns is because each individual grasshopper or mealybug egg is counted as an individual. The result is to greatly increase the apparent total animal population in the fall and early spring. But when each cluster of eggs is counted as one individual, the Orthoptera become negligible and the Homoptera decrease in numbers from early fall, the resulting total animal population showing a very striking uniformity throughout each season.

The one third increase in population in the spring over that in the autumn is due, not to insects, but almost entirely to earthworms and sowbugs-the earthworms being most abundant early in the spring when the ground is moist,

and going deeper than six inches as it later dries out, but the sowbugs become most abundant in May. Variations in the Hymenoptera column are due to occasional accidental selection for examination of a plot containing a nest, or near a nest of ants, but variations in the numbers of beetles, of which many species in greatly varying abundance were found, can not be assigned to any one cause. Considerable numbers of empty puparia were found, quite out of proportion to the small number of live Diptera. Thysanura were very abundant when weather conditions were just right, but Lepidoptera (mostly cutworms) and Hemiptera showed quite uniform abundance. The data as a whole show the preponderating abundance of earthworms, sowbugs, beetles, spiders and ants in this particular habitat. GEORGE N. WOLCOTT

1539 SUNSET AVE., UTICA, N. Y.

A COMMON-SENSE CALENDAR

OUR present Calendar is merely a rough tool. We can readily calculate dates inside of each month, and with some figuring we can reckon a month or two ahead. But when we are called upon to connect the proper weekday with some date more than two months ahead or behind, scarcely one of us can perform the operation without reference to a calendar. Current-year calendars are generally accessible, but it is no easy matter to ferret out old dates. It is even more difficult to determine dates in future years. In short, our measuring scale for dates is faulty. Like the Roman numerals it is unsuited for any but the simplest problems.

Who, for example, can find out for himself on what day of the week he was born? On what day of the week was the Declaration of Independence signed, or the Battle of Waterloo fought?

If your lease expires October 1 (or May 1), and you have to move, in what part of the week will this happen? If you have a regular engagement the first Monday of each month, will it conflict with another engagement on the third of next month or the month after? What months this year have five Sundays? How many annoying mistakes have you made during your life in such calculations?

If a ninety-day note or a three-month note is to be paid, on what day of the week is it due? Some quarters are longer than others, making the exact reckoning of interest difficult. Weekly periods of earnings in one year are not exactly comparable with the corresponding periods in other years. Holidays, like interest days, fall on different week-days in different years-sometimes very awkwardly. The lopsidedness of our calendar is due to the Emperor Augustus, who insisted that his month should contain as many days as the month of Julius Cæsar. As a matter of his