as though we are justified in thinking of the ear as whole as a coupled vibratory system and referring to the maxima of sensitivity as its resonant frequencies. The resonant frequencies are not to be thought of as harmonics of the fundamental.

Sensitivity of Abnormal Ears.-In the study of abnormal ears, where disease has affected certain portions of them and has produced partial deafness, curves of two types are plotted. The first type shows at the various frequencies how much current must pass through the receiver for the patient to just hear the tones compared with that required for a normal ear to hear. The second type shows the sensitivity of the ear at the various frequencies. Both methods of plotting the data are of value for studying the physical characteristics of the ear.

In figures 2 and 3 are shown curves relative to the normal reference ear for two totally different types of deafness. The former curve is for a case of oto-sclerosis (so-called "Fixation of the Stapes"), a type of middle ear deafness and the latter curve is for a case of internal ear trouble (so-called "Nerve Deafness"). The curves are alike in no respects, whatsoever. In the case of middle ear trouble there are marked depressions relative to the normal reference ear (figure 1) in the regions of 600 to 1200, 1500 to 2000, 2000 to 2500 and 3000 to 3500 cycles. There have been taken for different patients with middle ear deafness several curves and all of them agree in the location of these depressions. Some of these depressions are relative and some of them are absolute as reference to figure 4 will show. This figure (4) gives the sensitivity at the various frequencies of the ear referred to in figure 2. The two curves shown in figures 2 and 4 together give one accurate information on the physical characteristics of the defect. The stapes is not fixed but simply hindered and modified in its motion so as to produce less sensitivity than is possessed by a normal ear and also so as to give a sensitivity curve with three distinct resonant frequencies all of which are at higher frequencies than the corresponding ones for the normal reference ear and, as is to be expected, the peaks at the lower frequencies are displaced toward higher frequencies more than those in the upper part of the scale.

The curves for the case of the internal ear trouble shown in figures 3 and 5 are of much value in suggesting how the internal ear system functions physically. The patient's hearing is essentially the same as a normal ear up to about 2600 cycles as shown by figures 3 and 5; shown better in the latter, however. Then, the depression (shown better in figure 3) from 2600 to 4500 cycles is very great. When one recalls that the vibrational energy of the receiver diaphragm is proportional to the square of the receiver current, he will then grasp the significance of these curves in showing the enormous increase in the vibrational energy necessary for the patient to just hear compared with a person possessing normal hearing.

To show the ability of the person to check the data the dotted curves, taken six months later, are shown in figures 3 and 5. The curves not only check the earlier data but also show in the best possible manner that there has been no change to speak of during the six months' period. Several curves of this type have been obtained for different patients, but with the depressions showing up at different frequencies. We may ask how is it possible for so serious a defect to affect the hearing in only a single region of this sort? Without going into the anatomical evidence regarding the structure of the ear, we may say that physically the affected parts in these cases are evidently very loosely coupled (speaking mechanically) to the rest of the ear which functions normally out side of this region. Such would not be the case for middle ear trouble as shown by figures 2 and 4. The limit of audibility for this patient, who is 25 years old, is about 12000 cycles which is lower than we should expect for a person of this age. So that, the whole region above 3000 cycles (roughly speaking) is probably affected. From arguments to be presented elsewhere? it may be that this lesion is one affecting the transverse fibres of the basilar membrane. The data at hand seem to indicate that these fibres function in somewhat the same way as postulated by Helmholtz and others (not however as pitch "determinators.") Hence, possibly the fibres above 3000 cycles are prevented from performing their proper function while those below this frequency are not yet affected.

Auditory Nerve Endings.-In concluding this brief account it is worth while to show how the data suggest the part played by the nerve endings, not as organs for changing mechanical energy into auditory nerve energy, but as organs which respond only to a very narrow range of frequencies where the pitch discrimination is good and which respond to wider ranges where the discrimination is not so good. Some curves were taken which showed that within a range of one hundred cycles the hearing decreased suddenly so that the patients were unable to hear even though the vibrational energy of the receiver diaphragm was made 10,000 times as great as was necessary for them to hear at the lower frequency and perhaps 1010 as large as was required for a normal ear to hear at the frequency in question. With this amount of energy the region of the internal ear corresponding to the lower frequency was certainly vibrating with an amplitude considerably in excess of that necessary for audibility. Yet, no tone was heard. In other words, apparently it would be necessary to stimulate the nerve corresponding to the higher frequency if the tone of the higher pitch is to be heard. So that, apparently a nerve will respond only to one frequency or a very narrow band of frequencies. The actual pitch determination, then, resides in the nerve endings which appear to have this highly selective action. It seems probable, then, that this work is leading into fields which can be explored by precise

physical measurements and fields which up to the present time have not been covered satisfactorily by other means of investigation.

The writer is greatly indebted to Dr. J. Gordon Wilson, Head of the Department of Otology Northwestern Medical School, with whom this co-operative research has been done and to whom equal credit for the success of the work is due. The writer is also indebted to Professors Millikan and Lunn of the Department of Physics of the University of Chicago for their helpful suggestions and enthusiastic interest during all the stages of development up to the present time.

1 John P. Minton and J. Gordon Wilson, "Sensitivity of Normal and Defective Ears for Tones of Various Frequencies," Proc. Inst. of Medicine, Chicago, 1921.

2 John P. Minton, "Physical Characteristics of the Ear," in preparation for the Physical Review.

NON-DISJUNCTION AND THE CHROMOSOME RELATIONSHIPS OF DROSOPHILA WILLISTONI

BY REBECCA C. LANCEFIELD AND CHARLES W. METZ

STATION FOR EXPERIMENTAL EVOLUTION, COLD SPRING HARBOR, N. Y.

Communicated by C. B. Davenport, May 5, 1921

Of the eleven types of chromosome groups found in the genus Drosophila (Metz '16)' that designated type A is the most widespread-occurring in 13 of the 29 species that were studied. This type is represented in figure 1. It consists of one pair of rod-like chromosomes, two pairs of long, V-shaped chromosomes and (usually) one pair of very small spherical "m" chromosomes. The constancy of this type among the 13 species is such as to suggest a genetic homology of the respective pairs of chromosomes throughout. On the other hand the species themselves are scattered more or less at random through the genus and do not constitute a restricted taxonomic group. This suggested the desirability of a comparison of the genetic constitution of the chromosomes in two or more of these species.

Since D. melanogaster was already well known genetically and cytologically, and thus afforded a convenient basis of comparison, we undertook a study of D. willistoni, a species resembling melanogaster in appearan e and in chromosome constitution. In the course of this study we have found, by an examination of non-disjunctional flies, that the chromosomal resemblance between willistoni and melanogaster is misleading, for the sex chromosome pair of willistoni does not correspond morphologically to that of melanogaster but corresponds, rather, to one of the autosome pairs of this species. This would indicate that corresponding chromosomes in the two species are not all strictly homologous. The evidence for this conclusion together with a brief discussion is given below.

Figures 1 and 2 are diagrams; figures 3-10 are camera lucida drawings from fixed and sectioned material. Figure 1. Chromosome group of Drosophila melanogaster female; X chromosomes represented in solid black. Figure 2. Chromosomes group of Drosophila willistoni; X chromosomes in solid black. Figures 3 and 4. Normal female groups of willistoni. Figures 5 and 6. Normal male groups of willistoni. Figures 7-10. XXY groups from non-disjunctional females of willistoni.

Cytological Data.-The normal female chromosome group of Drosophila melanogaster is represented diagrammatically in figure 1 and that of D. willistoni in figures 2-4. In melanogaster, as shown by Bridges® ('16) in his work on non-disjunction, the rod-like chromosomes are the X chromosomes. These are represented in solid black in the diagrams. In the male of melanogaster this pair is asymmetrical, X being rod-like and Y J-shaped. In the male of willistoni no asymmetry is evident in any pair (figs. 5, 6). Otherwise the chromosomes of willistoni correspond, member for member, with those of melanogaster, except for the possible absence of the minute, spherical pair.

The lack of inequality in the sex chromosomes of the willistoni male necessitated obtaining non-disjunctional flies with three sex chromosomes in order to distinguish sex chromosomes from autosomes. As shown by the genetic data given below these non-disjunctional flies were females having one Y and two X chromosomes. Four chromosome groups from such females are shown in figures 7-10. In each of these there is an extra chromosome, and it is clearly a large and V-shaped one. Since this must be a sex chromosome it proves that the sex chromosome pair in willistoni is one of the long, V-shaped pairs and not the rod-like pair as would have been expected from analogy with melanogaster. We have obtained numerous clear cut figures both of the normal group in each sex and of the non-disjunctional group, and are confident that the evidence is entirely conclusive on this point.

Genetic Data. In obtaining and identifying the non-disjunctional flies we have followed the procedure used by Bridges ('16) in the case of D. melanogaster. The following summarized account shows the resemblance between the results in the two species and indicates the source of the XXY chromosome groups described above.

Primary non-disjunction appears to be very rare in willistoni and no cases (which could be checked) were observed in our regular experiments. In approximately 150 cultures made especially for this purpose only two cases were detected. In each of these the exceptional fly was a female. The first came from a cross of a rough eyed female by an orange, smallbristle male, both from stock cultures. (Rough, orange, and small-bristle are all inherited as sex-linked recessives.) This mating gave 132 normal daughters, 111 rough eyed sons, and 1 rough eyed (exceptional) daughter (Culture W 1715). Since this rough daughter had already mated with a rough brother, it was not possible to detect "secondary" exceptions produced by her. Six of her rough daughters, however, produced 11 exceptional sons and daughters among 960 significant flies, or 1.14 percent of exceptions. In succeeding generations about the same ration was maintained, as is shown in table 1. For the sake of convenience, this strain has been designated as "line A."

The second primary exception appeared in an entirely unrelated stock, involving the characters two-bristle, short-3, and rough. It gave rise to a second strain known as "line B." Table 1 gives a summary of the breeding tests with this line, which gave results essentially similar to those of line A.

In addition to the ordinary cases of "reductional" non-disjunction we have also detected two cases of "equational" non-disjunction. This type was first described by Bridges ('16) who reported 18 "equational exceptions," i.e., daughters homozygous with respect to a recessive character for which the mother was only heterozygous. In most of his cases, he