Carnegie Institution of Washington. Publication no. 399. (1929)
VARIABILITY OF EYELESS
By T. H. MORGAN

SELECTION OF EYELESS

Eyeless is a mutant whose gene is in the small fourth chromosome. Several stocks are kept that arose independently. Either they have allelomorphic genes for eyeless, or their differences may be due to modifying factors. One, the original eyeless, was found by Miss M. A. Hogue in 1914; the other, eyeless 2, was found independently by Dr. Nonidez and the writer, in the spring of 1919 in the same stock. Another, eyeless 4, was found by Dr. Li (1925). In all the original stocks the eyes were very variable as indicated in some of our earlier figures of eyeless (Bibliographia. Genetica II). In some individuals both eyes are present and nearly full size; in others both may be small, equal or unequal in size; in some individuals one eye is present (large or small), the other absent. In still others both eyes are absent. The three ocelli are generally present, although often—especially where one or both compound-eyes are absent—one or more may be displaced, or extra ocelli may be present.

The stocks are now kept with little selection, and remain at about what may be called the upper level, i.e., flies having both eyes, one eye, or no eyes. When both eyes are present they are, as a rule, smaller than normal, but often approach the normal so nearly that they cannot be separated and even overlap normal. The eyes are often very unequal on the two sides.

METHOD OF SELECTION

In general, two methods of selection were followed. In each generation Pair-matings were made and the most extreme individuals were selected as the parents of the next generation. Such a procedure is best if the results of selection are to be followed carefully in each generation. On the other hand, those flies that are without compound-eyes are blind, and pair matings are not so successful; for, the individuals may not mate for so long a time that the food becomes mouldy and the flies themselves get entangled in the mould. But if in each generation, small mass cultures of 6 to 12 flies are used, choosing as far as possible only virgin females, the chance that the culture will succeed is far greater. It is also a time-saving procedure. The determination as to what has taken place in the stock can then be attempted after the selection has gone so far that no further changes are apparent.

In order to find out when selection had ceased to bring about further modification, some method had to be devised to give a measure of each successive generation. The difficulty was to find some convenient way to measure the eyes. No doubt, measurements of the eyes would give the most accurate data, but this is too laborious, and technically too difficult, as also would be an attempt to count the ommatidia. I resorted to the simpler method of recording whether each fly had no eyes at all, or one eye, or both eyes, regardless as to whether the eye, when present, was large or small. [142]

It soon became apparent that any accurate measure was out of the question, since each culture undergoes a change from day to day. As the culture becomes older, the number of flies with one eye or with both eyes steadily increases in number. In extreme cases none of the flies in a given culture had eyes at the first count, while on the tenth day all the flies might have two eyes, or one at least. Since the cultures do not all go at the same rate, it is difficult to institute comparisons between them except in a very general way.

The change in the character of the offspring is not due to an age difference in the parent; for, if the old flies are taken out after 9 or 10 days, and put into fresh bottles, their first offspring are the same kind in the second bottle as those that had been produced in the first bottle.

It might be supposed that the eyeless flies hatched first and those with eyes later, but that this is not the explanation of the change was shown by moving the old flies from day to day to new bottles when all the offspring that hatched were like the first counts of a bottle.

CHARACTER OF EYELESS

When the eye is entirely absent the area that it would usually occupy is covered by a smooth cuticle. Since this area does not bulge out as does the normal eye, the head has, from side to side, a laterally contracted appearance as seen from above. The absence of the outer optic ganglia may also contribute to this reduction in size. The condition of the brain and ganglia has been described by Mrs. Richards.

When a small eye is present, it sometimes occupies the center of the optic area, and is surrounded by a cuticular zone representing the missing part, but it may lie in an eccentric position. The small eye is as a rule circular or slightly oval in outline, sharply bounded, and composed of ommatidia approximately normal in size, but fewer in number than normal. In extreme cases, the eye has only two or three ommatidia, and sometimes only a splash of red pigment is present.

ANTENNAE IN EYE

The development of an antenna in the eyeless region is of frequent occur­rence. Some duplicated antennas have been figured in Bibliographia Genetica, 1925, Vol. II, 72. The antenna, if in the eye region, has nearly always the same orientation as that of the normal antenna of the same side. In other words, it is not a mirror image or duplication of the normal antenna of the same side, but an organ, in series, equivalent structurally to a normal antenna. On the other hand, the antenna may appear in duplicate and be mirror figures. These are always near together and arise as a rule from a single basal joint. It is not without interest to note that the antenna that replaces the eye is usually quite perfect in structure and size. It may be recalled that in certain crabs, Herbst has shown that an antenna develops if the eye and optic ganglion are removed. The physiological significance of these relations is obscure.

According to recent work by Chen on the imaginal disc of the eyeless Drosophila the antennal part of the cephalic imaginal discs is present and of [143] normal size even when the eye part of the common disc is almost completely absent. It may be that the absence of the eye portion of the disc has a close relation to the doubling of the antennal portion, but just what this means is not apparent. It might mean more space for the development of the antennal part which causes it to double, or some difference in the shape of the antennal part leads to duplication. On the other hand, it may equally well be supposed that the extra antennae come from the rudiment of the optic portion of the disc when it is present.

TROPISMS OF EYELESS

The adult normal-eyed Drosophila is strongly positive in its light reaction, both when it crawls and when it flies. At rest, however, it is almost indifferent and may even stand turned away from light. The flies without compound eyes show no reaction to light, although the ocelli are present. It follows that the ocelli of the flies have no significance in normal orientation, and also that the light does not act directly on the brain, or at least not on that of the eyeless flies. It would be hazardous to extend this conclusion to other insects, for it is possible, though unlikely, that the ocelli of eyeless are affected by the other conditions present in the flies. When the flies have small eyes on both sides they show a positive reaction, but whether it is as quick or as accurate as that of normal-eyed flies has not been made out.

When the flies have a large eye, or even a small eye, on one side only, they show circus movements as they crawl, whether illuminated from one side or continuously from above. If illuminated from one side, the fly, if starting to the light turns to the eye-side and, as it does so, tends to place the eye in the shadow of its own head. This, however, stops the turning unless enough reflected light is present to cause a continuation of the circling until the eye emerges on the opposite side into the light again, etc. When illuminated from above the circling is due to a light continually impinging only on one side-there is no movement here towards the light, but one towards the side that is affected by the light through the eye.

When a fly has one large and one small eye, it often shows circus movement towards the large eye.

PROGRESS OF SELECTION FOR NO EYES

The experiment began in January 1921, at Palo Alto, California, by selecting a few flies from stock that had small eyes. In the second generation the selection was reversed, i.e., selection of big eyed flies took place. The following selections were "downward" again. Definite records of the condition of the eyes began to be gathered only in the seventh generation of the records at this time, as seen in the Appendix, are as extreme as those of today (1926), but others are by no means so good as at a later date. This means that the genetic possibilities for extreme eyeless were already present in the seventh generation, and probably, therefore, in the stock at the start. That the genetic factors that assist in producing a larger number of flies without. eyes in the early counts had not been at the time segregated, is shown by selecting as parents the flies that had the two largest eyes. The response was immediate, giving a greatly increased number of flies with both eyes, and relatively fewer with one or with no eyes. [144]

As stated above, the first requisite for a study of the effects of selection on the eyeless character was to find some approximate objective measure of the condition of the eyes. The flies were sorted out into three classes, viz.: (1) without eyes (2) with one eye (3) with both eyes. Experience showed that when there were many flies with no eyes in a culture, those with only one eye were few, and those with two eyes scarcely present at all. Conversely, if many flies have both eyes, there are many with one eye, and very few with no eyes.

The classification, unfortunately, does not express another relation, namely, that when many flies in a culture are without eyes, those with one or with two eyes have as a rule much smaller eyes than the single-eyed and two-eyed individuals of the original stock. In order to show this relation a rough estimate of size of the eyes was taken in 1922. The data are given in tables 1, 2, 3. It is evident that there is a high correlation between the total absence of one eye and the size of the other eye when present.

TABLE 1—Unselected Eyeless Stock

L>R L+R L<R
44 145 58

TABLE 1A—Summary of Table 1

3/4 1/2 1/4 1/8 0
R L R L R L R L R L
167 159 48 54 19 20 5 5 8 9

It has recently been shown by Chen that the imaginal eye-disc of the completely eyeless-fly is almost totally absent. Furthermore the larvae of the late hatching "eyeless" flies in which the eyes begin to appear, show corresponding changes in the amount of development of the eye disc, ever, at an early stage. Presumably when a difference appears on the two sides of the same fly, this difference also is present in the discs although this is not reported by Chen.

In order to find out something about the difference in size of the eyes in the same fly, a classification of the condition of the eyes in several stocks was made. In each fly, the right and the left eye was classified as full eye, 3/4 eye, 1/2 eye, 1/4 eye, 1/8 eye, and no eye. After these data had been collected, the classes were sorted out according to size. When both eyes were present each was put into its own class; when only one eye was present (the other absent) it was of course recorded once on its proper side. The [145] condition of the control eyeless stock at this time ("Woods Hole") is shown in table 1. Nearly all the flies have two eyes that are 3/4 or 1/2 full size.

A summary of table 1 is given in table 1a, in which all the eyes that are 3/4 or 1/2, etc., are classified according as to whether each is on the right or the left side. There is no observable tendency for the eyes to be larger on one side (right) than on the other (left). Both eyes were recorded as of the same size in 145 flies and unequal in 102 flies.

TABLE 2—Select Up Eyeless 20th Generation

L
R
3/4
3/4
3/4
1/4
3/4
0
0
3/4
1/2
1/2
1/2
1/4
1/4
1/2
1/2
0
0
1/2
1/8
1/2
1/4
1/4
1/4
0
0
1/4
1/2
1/8
No. 91 3 7 1 4 11 10 2 1 2 1 54 1 1


L>R L+R L<R
20 147 22

TABLE 2A—Summary of Table 2

1 3/4 1/2 1/4 1/8 3/8
R L R L R L R L R L R L
65 67 112 107 7 10 3 4 2 0 0 1

One eyeless stock had been selected through 20 generations for large eyes ("select up"). Its condition is recorded in table 2. All the flies had two eyes and nearly all were approximately full size or 3/4 size. Again, there is no asymmetry observable (table 2a). Both eyes were recorded as the same in size in 147 cases, and as different in only 42.

TABLE 3—Select Down Eyeless 23d and 24th Generations

L
R
3/4
3/4
3/4
1/4
3/4
0
0
3/4
1/2
1/2
1/2
1/4
1/4
1/2
1/2
0
0
1/2
1/8
1/2
1/4
1/4
1/4
0
0
1/4
1/2
1/8
1/8
0
1/4
1/2
0
1/8
No. 1 1 3 6 7 5 1 18 20 1 4 15 26 1 3 1 4

L>R L=R L<R
46 12 69

TABLE 3A—Summary of Table 3

1 3/4 1/2 1/4 1/8 0
R L R L R L R L R L R L
0 0 7 5 30 31 36 21 5 4 39 56

[146] [147] [148]

Another stock derived from the same original stock had been "selected down" for 23 and 24 generations. Its condition at the time is recorded in table 3. Here most of the flies have eyes only 1/2 and 1/4 the full size. There is no asymmetrical distribution observable on the two sides (table 3a). The eyes on the two sides of the same fly were recorded as equal in only 24 cases, and as different in 209 cases. The difference is so large that it cannot be regarded as due to greater ease of observing the differences that exist, but indicates rather that there is a much greater variability in the size of the smaller eyes. A very considerable number of these flies lacks one eye completely. This occurred 78 times on the right side and 112 times on the left side. It seems doubtful if this difference is indicative of an asymmetrical relation.

Another consideration shows that it would not be worth while to spend much time making accurate measurements of the eyes. As the culture gets older from day to day the character of the population generally changes, depending on the kind of alterations that take place in the food. A characteristic change is illustrated in the three counts recorded in table 4, each from a single bottle, at intervals of about two days. These counts show that while at first nearly all the flies are without eyes, there is a shift until nearly all have both or one eye at least. On the other hand, if the fermentation goes well and the bottle remains moist there may be no changing-over during the first seven or eight days, as the four counts in table 5 show. By adding new food (fermenting bananas) every few days the change may be to a large extent prevented as shown by the six records in table 6. The days on which the counts were made are given in the first column.

It has been pointed out that this shift is not due to the increase in age of the parents, for if transferred to a new bottle their offspring repeat the story. The result is due most probably either to the amount of food or to some other change in the environment.

It is not a little curious to find that the best fed and biggest flies are those less likely to develop eyes, while the late hatched small starved flies have large eyes.

PROGRESS OF SELECTION "UPWARD" FOR FULL EYE

A few flies whose eyes were not quite full were picked out from stock and bred together. From their offspring and in subsequent generations, flies with eyes as nearly full as possible were picked out to start each new generation. Almost at once the effect of selection was evident. In the seventh and eighth generations, after selecting "upwards" a census was made with the results given in table 7.

CROSSES BETWEEN STOCKS

Crossew were made between the up and down (+ and -) selected stocks in the seventh, eighth and ninth generations. In each case the flies were taken from the most extreme bottles of their kind. The results are given in table 8. [149]

It is quite evident that the more nearly full-eyed (plus) stock is, in a sense, the dominant. It follows that for testing the F1 flies in order to find out if the difference in the two stocks is due to modifying factors, the "select down" eyeless (as the more recessive stock) should give more valuable results.

TABLE 7

  Both One side
fleck of eye
One side
eyeless
Select up 7 247
120
74
157
0
0
0
0
2
0
0
0
Select up 8 278
127
154
119
51
417
2
3
2
60
0
1
2
0
0
12
0
0
Select up 10 47
66
0
0
0
0
Select up 11 160
173
0
1
0
0

  TABLE 8

  Both One None
Select down 7 by select up 7 24 9 (Intermediate)  
Select down 7 by select up 7 84 (All big to intermediate)  
Select down 8 by select up 7 26 14 7
Select down 8 by select up 9 138 19 3
Select down 8 by select up 9 15 2 0
Select down 8 by select up 9 46 2 0
Select down 9 by select up 9 19 16 2
Select down 9 by select up 9 284 46 9
Select down 9 by select up 9 27 7 0
Select down 9 by select up 9 28 0 0

The difference in the plus and minus stocks present in the 13th to 17th generations of selection is to some extent brought out by the following experiments in which flies from each stock were first crossed to Curly Hairless stock and the F1 back crossed to either the plus or minus selected stock. For example, select-down "eyeless" flies (14th and 17th generation) were mated to Curly Hairless. The F1 Curly Hairless males were bred to selectdown flies. The results are recorded in table 9.

In another case select-up flies (14th, 16th, and 17th generations) were bred to Curly Hairless, and the Curly Hairless F1 males were back-crossed to select-down (16th generation). The results are shown in table 10. [150]

TABLE 9

The difference in the two cases is not conspicuous. On the assumption that the Curly Hairless stock brings in modifiers for eyeless, especially those in the plus direction, their redistribution in the second generation would tend to cover up the recessives in that generation. On the other hand, when the select-up flies (13th and 14th generations) were bred to Curly Hairless and the F1 males were back-crossed to select-up flies (15th generation) the results are somewhat different from those of table 10 as shown in table 11.

Normal eyes Eyeless
Both eyes One eye None
(1) 154 154 57 2
  240 25 7 0
(2) 230 39 1 0
  285 11 0 0
  160 20 1 0
  250 25 0 0

  TABLE 10

Normal eyes Eyeless
Both One None
5 2 0 0
44 11 2 0
64 24 9 0
60 17 15 2
126 24 8 2
17 10 7 1

  TABLE 11

Normal eyes Eyeless
Both One None
114 4 0 0
169 29 0 0
41 8 0 0
31 5 0 0
41 3 0 0
15 2 0 0
176 41 3 0

Here all except three of the F2 eyeless flies had both eyes present. The result may be interpreted to mean that the Curly Hairless brought in [151] plus modifiers which were similar to or the same as those in the select-up stock. A census of the select-up stock in the 18th generation gave 350 flies with both eyes, one fly with one eye and none with no eyes. The outcrossing bad then virtually no effect on the condition of the eyes in the select-up eyeless stock. By implication, the select-down stock owes its difference to the accumulation of recessive factors that tend to obliterate the eyes.

LOCATION OF THE MODIFYING FACTORS FOR EXTREME EYELESS

The usual method for locating factors by the use of dominants in the second and third chromosomes was applied to eyeless. The first tests were made when the selection had gone only as far as the 7th and 8th generation of selection. The results were not wholly satisfactory, indicating perhaps that the stock was not yet homozygous, and the experiments were repeated later. Nevertheless a few examples may be given that will serve as an introduction to the difficulties that this case presents.

An eyeless female of the "select down" stock in the seventh generation was bred to a Curly male—a dominant second chromosome mutant. All the offspring had full-sized eyes. The F1 males that were Curly were backcrossed to females of the select down stock in the seventh or eighth generation. The F2 (back-cross) offspring are given in table 12.

TABLE 12

F1 Cy (Cy by down 7) by down 7 N Cy ey Cy ey Both One None
9 8 4 5 2 2 5
F1 Cy (Cy by down 7) by down 8 26 20 25 20 16 18 9
F1 Cy (Cy by down 7) by down 8 13 20 0 9 2 6  
48 48 29 34      

The flies were first separated into the eyeless and not-eyeless groups, and then each group subdivided into Curly and not-Curly (i.e., normal wings N). In addition, in each cross, samples of the eyeless flies were classified according to whether both eyes, one eye, and no eyes were present.

If the eyeless flies were as viable as the normals, equal numbers of these two classes are expected. In fact, however, there were 96 normals and 63 eyeless, but since it is known that some of the eyeless flies have full eyes (or eyes that appear to be full size) the excess of normals may be due, in Part, to their inclusion in this group. On the supposition that the presence of modifying factors increases the percentage of flies without any eyes, it is to be expected that in those F2 classes of eyeless in which the original chromosome complex has been recovered there will be a higher percentage of flies with no eyes than in those classes with a new combination of chromosomes, which here is the class with the Curly chromosome. Unfortunately, this subdivision was not made at this time in such a way that this can be shown, but it was made in later experiments. There is, however, another criterion that may be applied which is not so good. If it is assumed [152] that the modifier, that makes for more flies with no eyes, also increases the chance that in the F2 generation there will be a higher percentage of eyeless,—i.e., fewer of the genotype, "eyeless" flies having full eyes, then when the original complex of chromosomes is recovered (as in the not-Curly, eyeless class) there will be more eyeless flies than in the class that contains the foreign chromosome, viz., Curly. As a matter of fact, no marked difference is observed and such as it is (the difference present is probably not significant), it is in the opposite direction.

The preceding discussion rests on the assumption that the modifier is recessive. If it is dominant no such differences will be expected in the F2 generation.

These results, as far as they go, furnish at least no evidence that there are recessive modifying factors in the second chromosome. A similar test was made for the third chromosome using the dominant factor, Hairless. One experiment gave aberrant results and is discarded. The other is given in table 13.

TABLE 13

  N H ey ey H Both One None
F1 S H (H by select down 7) by down 7 25 25 14 16 12 7 11

The results are similar to those for Curly and do not furnish evidence of a recessive modifier in the third chromosome.

Since these two experiments were made with different stocks, namely Curly and Hairless, they are open to the objection that each stock may have introduced different modifiers that affect the result. In order to avoid this, the experiment was repeated using both dominants at the same time. Curly Hairless males were bred to selected-down eyeless females in the seventh and eighth generation, and the F1 Curly Hairless males were bred to selected-down eyeless females of the ninth and tenth generation. The results are shown in table 14.

TABLE 14

 

H ey H Cy H ey Cy H N Cy ey ey Cy Eyeless
Both One None
F1 CH (Cy H by down 7) by down 9 10 3 12 7 10 18 1 4 4 9 8
F1 CH (Cy H by down 7) by down 10 17 4 4 6 10 20 6 8 9 9 3
5 0 3 2 0 1 1 2 6 0 0
F1 CH (Cy H by down 8) by down 10 12 2 22 11 10 12 1 6 9 9 3
44 9 41 26 30 51 9 20 28 27 14

Only about one-third of the flies are eyeless (166:64). Comparing class by class the eyeless with the not eyeless, there is no consistent evidence showing that the percentage of eyeless is higher when the original second and third chromosomes are present than when one or the other is derived [153] from a foreign source, namely, the Curly and Dichaete chromosomes. This is apparent in the following tabulation.

When Hairless is present = 44 full 9 eyeless
When Curly is present = 51 full 20 eyeless
When Hairless and Curly are present = 41 full 26 eyeless
When neither is present = 26 full 9 eyeless

Owing to the environmental influence on the condition of the eye, it seemed very questionable whether the foregoing differences have any genetic significance, especially since they do not appear consistent amongst themselves. Therefore, an experiment was carried out on a larger scale with later generations (15th and 16th) when the homozygous conditions of the stocks was better assured. It did not seem necessary to classify the flies with full eye into their respective classes (as was done above) and the separation was made only of the flies with eyes smaller than full. In this material, one might expect to get evidence of the modifying factors, if present, that make more flies without any eyes. The results are given in table 15.

TABLE 15

Bottle Total Full eye Both eyes One eye No eyes
      N Cy H Cy H N Cy H Cy H N Cy H Cy H
0 37 22 2 0 1 1 0 1 1 1 1 3 1 3
ZB 40 25 6 3 0 1 3 1 1 0 0 0 0 0
AC 66 37 1 0 1 0 2 6 2 3 4 6 3 1
XX 76 39 1 0 6 3 0 8 0 6 7 6 0 0
AD 84 53 0 0 1 2 0 4 0 3 8 8 3 2
AE 91 54 0 1 5 1 1 2 2 5 1 14 0 5
PP 97 81 6 1 0 0 5 0 0 0 4 0 0 0
UU 98 81 8 2 0 0 3 2 0 0 1 0 0 1
E 128 68 0 0 3 3 1 6 5 6 10 9 6 11
ZZ 130 89 3 6 2 4 4 1 3 2 5 4 3 4
AB 133 32 6 6 2 3 2 12 10 13 13 13 10 11
QQ 145 69 0 2 4 6 7 6 4 8 10 11 3 15
WW 163 80 5 8 7 2 1 7 5 9 8 19 4 8
MM 193 138 4 2 2 1 2 4 4 6 14 7 3 6
BB 208 150 5 6 2 8 5 7 1 4 7 10 1 2
RR 240 139 0 6 11 10 10 7 11 5 11 8 8 16
LL 268 181 1 4 2 9 3 5 9 13 9 14 11 7
GG 297 178 0 6 11 11 6 11 6 12 22 18 6 10
YY 299 184 3 1 1 7 1 13 4 11 14 17 19 28
NN 319 309 7 2 1 0 0 0 0 0 0 0 0 0
OO 395 199 25 35 22 15 12 16 16 14 9 3 18 11
AA 396 327 1 2 4 6 7 5 6 6 18 6 2 6
HH 420 277 2 18 10 14 9 13 6 11 13 20 7 20
FF 428 350 12 6 5 2 17 5 3 5 6 5 8 4
88 504 336 3 3 9 13 7 20 6 16 21 19 25 26
  3794 116 121 113 124 115 166 106 163 216 214 139 203

The numbers of eyeless flies obtained would seem to be sufficiently large to show whether the second and third chromosomes carry the modifier in [154] question. Taken as a whole, the condition of the eyeless flies is somewhere between that of the original stock and that of the select-down stock. For example, these F2 eyeless flies if classified into three groups: with both eyes; one eye; no eye; respectively, give

Both eyes 474; One eye 550; No eye 772.

In comparison with the select-down stock of this generation, there are noticeably more flies with both eyes present. This disposes of the possibility that the gene for eyeless had been changed as a result of the selection.

Within the class, "no eyes," the class (216) with the original grouping of second and third chromosomes (= N in table 15) contained no more flies than the other classes, except the class (139) containing the third chromosome (H), but that this result is not significant is shown by the Cy H class which is not significantly lower than the N class, etc.

A similar analysis of the one-eye group leads to the same conclusion. Since there is little or no crossing over in the fourth chromosome-the one containing the eyeless gene-the redistribution of modifying factors for eyeless cannot be referred to this source.

There, remains only the X-chromosome. That the modifiers are not carried in the X-chromosome seemed determined by the way the experiment was carried out. The F1 male had received his X-chromosome from his eyeless mother of the selected-down stock. He was crossed to eyeless female of the same stock, hence all the second generation (back-crossed) flies had the X-chromosome of the selected-down stock.

The conclusion may seem to follow that the recessive modifiers sought for are not present in any of the chromosomes. It is true the Y-chromosome has not been considered. This came from the Curly Hairless stock, and is present in the F1 male, and transmitted by him to his sons. lithe effect were due to this chromosome, the F2 Sons should be more numerous in the no-eye group than the daughters. No such effect was observed. Moreover, if the effect were due to the Y-chromosome the male and female of the original select-down stock should be consistently different, but this is not the case.

The only conclusion that can be drawn is that the experiment was not made in such a way as to reveal the location of the factors present. This may mean no more than that environmental factors swamped the differences due to the genetic factors.

If the extreme eyeless condition (no eyes) is affected by specific dominant genes in either the second or third (or in both) chromosomes, these effects would not be brought out by the preceding tests, because the chromosomes containing them would be distributed at random to the F2 Curly and Hairless eyeless flies. If the double dominant, assuming it not to be lethal, gives a more pronounced minus effect than a single dominant, the expectation would be the same as for recessive factors. Since the not-Curly not-Hair- less eyeless flies were not distinctly different from the others, the experiments indicate that no such decided differences were produced by double dominants in comparison with single dominants. Moreover, crosses between select-up and select-down would be expected to show dominance in F1, but [155] the result was the opposite, namely, the select-up type was the more dominant.

It became necessary to show that some of these F2 flies could be again changed over to give the original distribution of the characters. This was done. As shown in table 16, for F3, F4, and F5, the extracted eyeless (from Cy H) were soon brought back to the extreme "eyeless" condition.

TABLE 16—Extracted eyeless (out of Cy H)

  Both One None
F3 0 7 60
F3 15 46 159
F4 0 4 46
0 3 44
0 2 12
0 4 22
3 9 26
9 9 23
F4 0 1 50
0 1 47
4 13 23
5 6 12
F4 0 1 19
0 2 40
1 8 38
1 9 13
2 0 7
F4 0 0 5
0 4 64
0 5 34
2 3 23
F5 0 2 9
1 8 25
F5 1 12 63
1 7 57
F5 0 2 65
0 7 62

TEST OF THE SELECT-UP STOCK

Experiments had shown that the ordinary stock responded quickly to selection in a plus direction. As a result, nearly all of the eyeless flies had two eyes present and these averaged larger than the two-eyed flies of the minus selected stock. Presumably, there were genetic factors present at first that modified the eyeless condition in a plus direction. In the first two experiments, the cross was made with Curly; in the second three, with Hairless. The results are given in tables 17 and 18.

There is no evidence from these results to indicate that recessive modifier are present in either of the chromosomes tested, but perhaps no answer 8 expected since the action of the introduced dominant modifying genes [156] would be to remove more "eyeless" flies into the full-eyed classes. There is no more obvious increase in the full-eyed (Normal) than in the full-eyed Hairless class.

TABLE 17

  N Cy ey ey Cy Both One None
F1 Cy male (Cy by up 7) by up 7
F1 Cy male (Cy by up 7) by up 7
34
41
26
44
23
33
20
27
19
142
2
3
0
0

 TABLE 18

F1 H (H by up 6) by up 7 N H ey ey H Both One None
48 45 43 31 38 2 0
F1 H (H by up 6) by up 7 17 33 10 20 28 2 0
F1 H (H by up 8) by up 8 75 111 39 17 31 1 0

In another experiment, the selected up eyeless flies (up9) were crossed to Curly Hairless, and the F1 male back-crossed to up" (table 19). The fact that in F2 the flies with the original complex (N and ey) were more nearly in a 1:1 ratio than were the other classes, may or may not be significant. The numbers are too small to be decisive.

TABLE 19

  H ey H Cy H ey Cy H N ey Cy ey Cy Both One None
F1 Cy H (Cy H by up9) by up9 10 2 13 6 2 1 5 6 15 1 0
F1 Cy H (Cy H by up9) by up9 26 5 30 4 14 16 21 8 33 0 0

Finally, in one experiment "select-down" flies were bred to the double dominant Curly Hairless (CyH) and the F1 males bred to "select up." The results (table 20) do not show any irregularities in the different classes (Cy and H, etc.). None of the flies lacked both eyes, demonstrating the dominance of factors in the select-up stock, or that the CyH stock itself carried plus modifiers.

TABLE 20

  H eyH CyH ey CyH N Cy ey eyCy Both One None
F1 Cy H (Cy H by down 7) by up8 15
15
15
2
5
10
35
23
33
0
6
8
23
23
24
9
32
11
11
13
10
3
6
13
16
24
42
0
5
2
0
0
0
F1 Cy Hmale (Cy H by down 7) by up9 12
22
11
15
16
20
9
12
33
66
20
24
21
13
6
2
12
20
8
19
50
53
18
17
5
8
5
15
12
23
9
13
42
54
11
12
0
F1 Cy Hmale (Cy H by down 8) by up8 33 11 50 11 35 38 19 12 42 11 0

[157]

SELECTION EXPERIMENTS

One of the objects of continuing the selection experiment in the minus direction, i.e., by selecting flies with no eyes, was to discover whether a stock entirely without eyes could be produced, and whether after many generations of selection, reversal of the selection would bring about any change. In regard to this first point it may be stated that after 60 generations of inbreeding the eyeless flies continued to produce some flies with eyes which were, as a rule, smaller than the full eye. Late hatching flies continued, as at first, to have eyes approaching more nearly the normal in size. This means that if the stock has become practically homozygous, either the gene for "eyeless" gives, as a phenotype, flies with small eyes as well as flies with no eyes and cannot be changed by selection, or else it means that the stock has become homozygous for genes that reduce the size of the eye which genes were probably present from the beginning. Here, as in all such experiments, it is difficult to discriminate between these two possibilities unless very elaborate linkage tests are made that involve all the chromosomes. It would be practically futile to undertake such work on a character of this kind that is subject to great variation in its phenotypic expression due to external conditions that are too complex to be fully controlled. Moreover the possibility of mutant modifying genes affecting so sensitive an organ appearing in the course of such an experiment would further increase the difficulties of the problem.

The second problem is much more feasible, and at several stages in the course of the work flies have been selected from the cultures that have both eyes or one eye or with no eyes to observe whether their offspring differ on the whole from those of the main line. Here again a peculiarity of the stock needs to be taken into consideration. In the earlier hatches there are only a few flies with both eyes present and these are usually small, while in the late hatched flies all may have two eyes and these may be quite large. Now if the former were more sensitive to the external conditions, they might or might not be the ones best suited for a test (depending on what was to be tested) or they might be those whose genetic make-up (if the stock were not entirely homozygous) favored the development of two eyes. In this respect, if the presence of genetic differences were being sought for, these might be better than the late-hatching flies. For, the late-hatching flies will obviously include flies with two eyes which involve all the genetic types Present, etc. It is obvious from these and other considerations that the Problem here is, as in all such cases, very complex. As a matter of fact, most of the selections in the positive direction were made from the earlier or middle hatchings. A few experiments were deliberately planned in which late-hatching flies were used. It may be stated that no measurable differences in the two cases was found.

Several selections towards full eye have been made. The earlier ones in the 9th, 16th, 17th, 18th, and 20th generations of select-down are given in the following records (table 21). There is no evidence in any of these cases that the first selection of the phenotype with one or two eyes present gives any difference in the offspring from the "no eyes" flies of the same stocks.

Later, a much more complete series of tests were made through three to [161] six consecutive selections (table 22). It is evident that no change was brought about by selection of the stock at this time. In other words, it had become practically homozygous for the modifiers of eyeless.

These data from flies in the 38th to the 51st generation are typical. In some cases the flies continued to be selected for one eye. The record in each case gives the last generation. In other cases, after one or another kind of selection (indicated in the table) had been obtained, reversed selection took place one or more times. Despite the rather wide variability there is no evidence that reversed selection produced any effect.

In addition to these selections, the records (table 23) carried several generations further than the last, may be added.

TABLE 23

  Gen. Both One None
No eyes for 6 generations, then one eye for 4 generations 54th 7 50 213
No eyes for 7 generations, then one eye for 4 generations 54th 1 11 65
No eyes for 8 generations, then one eye for 4 generations 54th 1 16 183
No eyes for 8 generations, then one eye for 4 generations 54th 9 45 276
No eyes for 8 generations, then one eye for 4 generations 54th 6 19 83
No eyes for 8 generations, then one eye for 4 generations 55th 0 18 134
No eyes for 8 generations, then one eye for 4 generations 55th 13 55 105
No eyes for 8 generations, then one eye for 4 generations 53rd 6 9 5
No eyes for 8 generations, then one eye for 4 generations 55th 7 30 62
No eyes for 8 generations, then one eye for 4 generations 55th. 22 67 128
No eyes for 9 generations, then one eye for 4 generations 56th 0 16 182
No eyes for 9 generations, then one eye for 4 generations 55th 1 34 93
No eyes for 9 generations, then one eye for 4 generations 56th 10 28 99
No eyes for 9 generations, then one eye for 4 generations 56th 0 9 10
No eyes for 9 generations, then one eye for 4 generations 55th 0 27 266
No eyes for 9 generations, then one eye for 4 generations 55th 4 16 121
No eyes for 9 generations, then one eye for 4 generations 55th 0 5 18
No eyes for 9 generations, then one eye for 4 generations 56th 0 5 127
No eyes for 9 generations, then one eye for 4 generations 56th 0 8 34
No eyes for 9 generations, then one eye for 4 generations 56th 8 31 139
No eyes for 10 generations, then one eye for 4 generations 57th 0 0 9
No eyes for 10 generations, then one eye for 4 generations 57th 5 27 155
No eyes for 10 generations, then one eye for 4 generations 56th 3 25 141
No eyes for 10 generations, then one eye for 4 generations 57th 5 26 215
No eyes for 10 generations, then one eye for 4 generations . 55th 0 13 157
No eyes for 10 generations, then one eye for 4 generations 56th 21 41 159
Both eyes for 1 gen., no eyes for 9 gen., one eye for 4 gen. 55th 0 0 12
Both eyes for 1 gen., no eyes for 9 gen., one eye for 4 gen.  55th 3 12 101
Both eyes for 1 gen., no eyes for 9 gen., one eye for 4 gen. 56th 7 20 21

In one experiment flies out of the same culture from the select-down stock In the 35th generation, of three kinds (three with two eyes; others with one eye, others with no eyes) were bred with the following results:

Parents both eyes 35th gen. Parents one eye 35th gen. Parents no eyes 35th gen.
Both One None Both One None Both One None
11 53 93 1 19 59 20 44 120

[162]

There is no significant difference here in the kinds of offspring produced by flies of the different phenotypes.

In the preceding data the end results of certain selections are given. As a control, the kinds of flies in the parental cultures were recorded, in some cases for one generation only (1), in other cases for several generations (2) (3) (4). A few samples of these are given to illustrate the fluctuations shown by successive generations. These records come from the end of the experiments, usually within the 50th to 60th generation (table 24).

TABLE 24—Three or more generations

No eyes One eye Both eyes
  Both One None   Both One None   Both One None
(1) 8 10 28   (1) 37 17 4 (1) 15 17 6
  14 21 63     11 35 42   5 10 28
(1) 37 17 4 (3rd count) (1) 6 17 4 (1) 25 2 2
  26 15 2     18 77 60   3 40 53
(1) 6 14 4 (3rd count) (1) 10 8 11 (1) 14 28 17
  4 4 4     2 5 14   6 5 23
(1) 13 4 4 (3rd count) (1) 5 27 22 (1) 2 24 14
  3 22 63     4 21 65   10 52 107
(1) 18 4 2 (3rd count) (1) 15 17 6 (1) 11 16 3
  22 23 28     1 1 18   1 2 25
(1) 2 10 15 (4th count) (1) 8 13 28 (1) 18 16 3
  6 13 48     8 37 59   0 4 14
(1) 5 27 22 (4th count) (1) 2 24 14 (1) 14 28 17
  22 8 0     10 52 107   6 5 23
(1) 2 10 15 (4th count) (1) 10 8 11 (1) 6 17 15
  1 24 99     2 5 14   65 105 88
(1) 1 17 8   (1) 5 27 22 (1) 67 67 62
  5 25 74     4 21 65   25 15 2
(1) 14 18 17   (1) 6 17 15        
  13 28 104     74 105 50        
(1) 2 24 14                  
  8 20 69                  
(1) 5 27 22                  
  26 36 64                  
(1) 2 24 14 (2nd count)                
  8 20 69                  
(1) 6 17 15                  
  24 61 61                  

[164]

THE NON-INHERITANCE OF AN ACQUIRED CHARACTER AS ILLUSTRATED BY EYELESS

In order that a character be "acquired" in the sense here employed it must be directly susceptible of being changed by some environmental agent, or indirectly affected by its use or disuse. It has been claimed that characters acquired in either of these ways may be or are transmitted through the germ-cells to the next generation. It might appear on first thought that those special characters that are most easily affected, or affected to a great extent would be the ones best suited to test the question of their inheritance. In fact, in most "experiments" in recent years it is this sort of material that has most often been used (salamanders and pupae). This is, however, a questionable assumption, and there is nothing known that supports such a view, but since this sort of character is the one most likely to attract attention, it has been often chosen to carry out such tests.

Modern genetic work has shown that the problem of the inheritance or non-inheritance of acquired characters is intimately bound up with the causes of variability in any given stock. It has been amply demonstrated that internal genetic factors may sometimes produce results that are superficially indistinguishable from those produced by external agents. The present case is an excellent illustration of this relation, for, both genetic factors and environmental factors change the eyeless flies either back to full eye or to entire absence of eyes. These possibilities must be watched, and, wherever they introduce disturbing factors, the experiment must be carried out so as to detect them. The most general rule is to make the material homozygous as far as possible, and in the second place to use pedigreed material whose genetic possibilities have been tested by inbreeding over several or many generations. An effort to do this has been made in the experiment with eyeless where the material was inbred for about 60 generations. The parents in almost every generation were without eyes, yet after about 60 generation of ancestors without eyes, the same kinds of flies were produced as at the beginning, viz., flies without eyes, with small eyes, and flies with larger eyes—as did the flies after the first three or four generations. The absence of eyes in the ancestors produced no change in the character of the offspring after the modifying genetic factors had once been made constant.

In the present case the flies bred from were kept without eyes or rudiments of eyes by the presence of genetic factors which were assisted by environmental agencies. This somatic absence produced no effect on the character; for, if it had done so, the eyes would either have disappeared, or have been further reduced as the selection continued. The data showed that this did not take place.

The evidence also shows that the principal gene was not changed, for whenever these flies were outbred and new eyeless stock subsequently recovered, it gave exactly the same results as did the original stock. The gene for eyeless was neither "contaminated" by outbreeding, nor was it changed by selection.