Portugaliae Acta Biologica - 1949
A Cytophysiological Theory of the Gene, Gene Mutation
and Position Effect

J. A. Serra

Secção de Zoologia-Anthropologia, Faculdade de Ciências. Universidade, Coimbra



* Matricial nucleoproteins are taken here in the sense referred to by Serra (1942, 1947). Kaufmann (1947) in an important review on heterochromatin has criticized our use of the old term matrix. We are perfectly aware of this term having been employed in a rather loose meaning by several workers, but as we defined precisely what we meant with matrix and are not fond of coining new names, consciously employed an old name giving to it a somewhat different conceptual context. If the old name matrix is to be dropped, then the true name of chromatin could be employed, since the matrix in reality corresponds to the chromatin proper.

† Polygenes is taken here in the assertion of factors, each with only a small action, of which several control a phenotype in a polymeric type of action. It must be pointed out that we do not consider the polygenes as exclusive of heterochromatin; of course, they must be present also in euchromatin.
Variegation following relocation to near chromocenter heterochromatin can not be explained by assuming repeated mutation of the gene: this hypothesis would be parallel to that which tried to explain position effect by a concomitant gene mutation, and is contrary to all the facts known about gene stability. The explanation which is more obvious is that something characteristic of heterochromatin, and itself variable, is responsible for this variability in effect—Fig. 3. Heterochromatin is known to possess during the resting stage an overcharge of matricial nucleoproteins (Serra 1942-1947) and this overcharge is variable, being greater in certain nuclei than in others, with a variability which has been studied in detail in certain cases (Fernandes and Serra 1944).* Moreover, has as been pointed out by Heitz (1929-1933) and many authors afterwards, there is reason to admit that this overcharge of matricial nucleoproteins is connected with the inactivity of the genes located in the heterochromatic region, as is shown by the following facts:

1) In Drosophila, normal genes are almost absent from the Y and the heterochromatic pericentromeric regions, though a "polygenic" activity may yet be present in these regions.†

2) It has been claimed that in other cases, v.g. in Hepaticae, the heterochromatic sex chromosome is not deprived of genetic activity (detailed review in Resende 1945). This, however, may be explained by admitting either that the genes are located in chromosome regions not heterochromatic (even in the most extreme cases of heterochromatic chromosomes it may be observed that there is always a part of the chromosome completely or almost completely despiralising and losing its matricial coat), or that the state of heterochromatization is different in other tissues, genic inertness being expected only in cells in which the chromosome region maintains its matricial overcharge during ontogenesis.

3) Excepting heterochromatic regions, the rest of the chromosome loses its charge of nucleoproteins in telophase to acquire it again only in the following prophase, while in interkinesis the charge is greatly reduced. It is safe to assume that the chief synthetic activities of the chromosomes take place during the so called resting stage, when the charge of chromatin over the chromosomes is at a minimum; conversely, during the condensed states of the chromosomes, during the middle stages of mitosis, the synthetic activities of the chromosomes must stop or be drastically reduced, as the layer of the chromonemata with the rest of the cell, a case in favour of this assumption being the well known loss of matrix during the growth of the oocyte in animals, after the chromosomes have attained diplotene. It follows that an over charge of matricial nucleoproteins corresponds to a low level of synthetic activity and this must be true also of the heterochromatic regions.

It seems a sound extrapolation to link both the inactivity caused by heterochromatization and the variability of the matricial overcharge of heterochromatic regions, with the inactivation of the + allele in heterozygotes R+/g (in which g is a mutant and R a rearrangement) and the variegation of the phenotype. Local development factors cause a variability in the matricial overcharge and this brings about a variegation, according to [whether] the full functioning of the + gene has been possible or not. It must be borne in mind that the non-functioning of the + gene corresponds to the manifestation of the recessive, at least in simple cases: if a v+ gene in Drosophila is hindered of producing triptophane-oxidase the result is the v phenotype. Also, hemizygous R+/- may show a more or less strong inactivation of the + allele, and concomitantly, a more or less typical mutant phenotype. In Drosophila melanogaster the inactivation of the + gene seems in many cases to be more marked in earlier development, giving rise to large patches of tissue of the mutant type in which there are scattered islands of wild tissue caused by reactivation in later stages of development (Demerec and col. 1939), as if the conditions which favour the overcharge of nucleoproteins realized chiefly in earlier development, which is a rather logical conclusion.

On the other hand, it is necessary to assume that the overcharge of nucleoproteins may extend to the neighboring segments and this in some cases over a length of about 50 salivary bands—Fig. 3. This idea has been expressed under a different form by Schultz (1936) who supposed that heterochromatic bands were characterised by a greater amount of nucleic acid. We have demonstrated that what deposits upon the chromonemata are nucleic acid and basic proteins (Serra 1943, Serra and Queiroz Lopes 1944, Serra 1947) which form a nucleoprotein gel, and therefore in heterochromatic regions an overcharge of nucleoproteins is to be found. However, the chromocenter of Drosophila in usual aceto-carmine or aceto-orcein preparations has a diffuse or "meshwork" aspect. This particular aspect of the chromocenter region of Drosophila has led to a great confusion about the interpretation of heterochromatin. In effect, it has been supposed that heterochromatic regions multiplicate at a slower rate than the rest of the chromonema, or that heterochromatin shows an allocycly in relation to euchromatin, that is the two kinds of chromatin would show a difference in staining (chromatic charge) at different cycles during the multiplication of the chromosomes (reviews in Resende 1945, Darlington 1947, Kaufmann 1947, Schultz 1947). It must be noted that not in all Diptera the heterochromatin is as that of the Drosophila chromocenter; for instance in Chironomus, at the two sides of the chromomere in the long salivary chromosomes there are regions believed with reason to be of heterochromatin which are overcondensed in relation to the rest of the chromosomes. And nevertheless, a meshwork aspect is not observed in these regions: only the chromosome ends show a fan-like structure, in part conditioned by the presence of more or less heterochromatin in the rest of the nucleus (Schultz 1947) and this fan-like aspect is also observed in the nucleolar region of the IV or nucleolar chromosome of Chironomus salivaries. It seems also that the "bulbs" and those parts of the salivaries in Drosophila with a low staining affinity and apparently less dense are also heterochromatic, for instance the "onions" at 2B, 11C, 16A, 25BC, 58E, 23C, 74, 75, 82F, etc.. Another important fact is that in Drosophila all transitions are found in different tissues from the condensed chromatin of the typically mitotic ganglionar chromosomes to the diffuse or meshwork chromocenter of the salivaries.

It must also be borne in mind that a lower rate of reproduction has been demonstrated only in the case of certain supranumerary or B-chromosomes, which show degenerescence changes: exagerated chromaticity and exagerated contraction, and a pycnotic aspect. For all the other cases, and especially for the heterochromatic regions or ordinary, chiefly euchromatic, chromosomes there is no valid indication that heterochromatic segments multiplicate at a slower rate than the adjacent euchromatic ones. It has been claimed (Darlington and LaCour 1940, Schultz 1941, 1947) that heterochromatin reproduces at a slower rate than euchromatin, but this is more an assumption derived from theoretical considerations than a conclusion proved by actual facts. A proof in favour of this assumption has been adduced from what happens in ovary nurse cells in Drosophila (Schultz 1941, 1947, Painter and Reindorp 1947) where the heterochromatic regions form about 4 or 8 strands or chromocenters, while the rest of the chromosome would make about 8 endodivisions (512 strands). However, this may be interpreted as signifying only that in heterochromatic regions the chromonemata tend to aggregate in bundles of 4 or 8 chromonemata, joined by the chromatic charge characteristic of heterochromatin.

Likewise, the hypochromaticity and lower diameter of the "starved" zones brought about by a differential cold reaction (Darlington and LaCour 1940) are clearly due to a lower amount of matricial nucleoproteins, and more particularly of nucleic acid, but this has nothing to do with chromonema reproduction since in normal cases similarly appearing zones, though with not so great length, can be seen in the nucleolar and anucleolar zones of normal chromosomes, for instance in Encephalartus (Resende and colls. 1945) and several other plants. The differential cold reaction is indeed an index of "affinity" of the zones to the matrix, and especially to its thymonucleic acid, but as far as can logically be inferred, the conclusion that in the starved zones there has been a slower rate of multiplication of the basic structures of the chromonemata does not seem legitimate.

The explanation of the peculiar aspect of the salivary chromocenter in Drosophila and other Diptera is to be sought for in the structure of the whole nucleus, and not in the multiplication of the chromonemata. First, it must be recalled that not in all the Diptera there exists heterochromatin of the meshwork type, and that in Drosophila itself all transitions from the compact to the diffuse heterochromatin are present. Second, there are other Diptera, for instance Sciara (see Metz 1941), in which the structure of the salivary chromosomes is a kind of network not far different from the meshwork of the Drosophila chromocenter, and the explanation to be admitted for this aspect is the imprisoning of the nucleoplasm or nuclear sap within the chromosomes, since the nucleus, without a nucleolus and almost devoid of nuclear sap, is almost all occupied by the chromosomes. Simply by pressuring the salivary cells in intact living Sciara larvae, or by a mild mechanical injury, as well as by asfixiation, it is easily possible to bring about a reversible shrinkage of the chromosomes, nuclear sap appearing then (Buck and Boche 1938). Those familiar with observing Dipteran salivary glands in vivo know that in many cases the state of hydration of the chromosomes is such that they are visible only with difficulty, and in such cases a good deal of nucleoplasm seems to lie within the chromosomes.

Third, it is also important in this connection to note that the peculiar aspect of the Drosophila chromocenter in aceto-carmine or aceto-orcein preparations is in part a consequence of the chemical action of acetic acid. In effect, we have observed that by a Feulgen coloration after an alcohol-acetic (3:1) fixation, the chromocenter of Drosophila does not show a meshwork structure. Its aspect is then not very different from that of the euchromatic bands, though in the chromocenter the bands are closer one anothers. Schultz (1947) by observation with contrast phase optics has also come to the conclusion that the meshwork aspect of the chromocenter is in part due to the action of the treatments. From these observations it follows that the nucleoplasm-containing chromocenter is only poorly fixable by acetic acid, the action of this chemical consisting chiefly in a gelification followed by a lengthy "precipitation" of the protein structures; those which have much water must not preserve very well with acetic acid. The probable conclusion is that the chromocenter has close relations to the nucleoplasm, this possibly signifying that the metabolism of this latter is affected by, or intimately related to, the heterochromatin; however, this is a point which only future work can clarify.

From all these facts, the right conclusion to be drawn is that to-date there is no convincing proof that heterochromatic regions of normal chromosomes (not of degenerating ones, of the B-type) do not divide at a slower rate than euchromatic regions. The hypotheses which endeavour to explain variegation by the loss of the + gene inserted near the chromocenter, due to a slower rate of reproduction of the heterochromatin and the near-by located + alleles, are also to be rejected on account of their lacking a sound basis. The observation that the mutant-type patches of tissue are produced in many cases preferentially in early development, is also against such hypotheses. What really happens is that chromocenter heterochromatin can influence euchromatic regions next to it by an heterochromatization process. In accordance with our hypothesis (Serra 1944), it is found that in reality the euchromatic segments inserted in the chromocenter are not lost; what happens, as is described by Prokofyeva-Belgovskaya (1947), is euchromatization or heterochromatization, which we discuss in detail in the following.



Allshire: RNAi, Heterochromatin (2002)

According to the hypothesis of an extension of heterochromatin over near-by relocated pieces, the euchromatic regions with the affected genes must show heterochromatization, that is a change of aspect towards heterochromatin. This change is, at least in principle, only one of aspect and not of nature, the euchromatic segment is more or less hindered from its normal functioning but its genes are not lost or mutated, since it is observed that reversion to the primitive location immediately brings about the loss of variegation and the position effect, that is a return to normal gene functioning. These cases of reversion are one of the best proofs against the hypotheses which assume that the genes, or their power of reduplication, are affected.

Conversely, according to the heterochromatization-euchromatization hypothesis it is to be expected that euchromatin also exerts an action, which may be called euchromatization, over adjacent heterochromatin, since it must hinder the heterochromatization—Fig 3. Euchromatic and heterochromatic regions exert, therefore, an interaction and the net result of the two contrary actions depends upon the "force" or potence of the two tendencies. Probably this relative potence depends upon the relative length of the two regions, their more or less pronounced heterochromatic nature and the point of the chromosome in which the relocation takes place. The influence of the relative length of the two regions, euchromatic and heterochromatic, which seems a simple postulate, may be obscured by the nature of the heterochromatic segment and other circumstances, for instance if it includes or not the nucleolar zone (Kaufmann 1943) which seems to be a strong heterochromatization factor. The influence of the other factor, the point where the relocation occurred, is yet more complicated since the true distribution of heterochromatic regions and of zones with heterochromatic tendency in Drosophila salivaries is not known with any degree of accuracy. The chromosome zones where nucleolus-like bodies with a nucleoplasmic aspect are seen must be taken as heterochromatic and the "bulbs" are also probably to range in this category. The incertitudes about this second factor cause that the determination of the amount of heterochromatization to expect in each case be a question to be settled only by actual experiment. When the distribution of the regions with heterochromatization tendency is accurately known, a more or less approximative prevision of the amount of variegation will be possible.

FIG. 3. Explanation of the position effects. Eu-Het, relocation of a piece of chromosome a (between the arrows) next to the heterochromatic part of another chromosome, b; an extension of the matrix charge (dotted) results, especially near the lower break, but also, in a lower degree, at the upper breakage point. The phenogenetically active zones or active nemameres are represented by groups of straight lines, each group forming a gene; while the inactivated nemameres are dotted. Het-Eu, a piece of an heterochromatic segment (between the arrows) of a, relocated in b, is euchromatized, only the breakage points showing matrix overcharge in a/b. Eu-Eu, an euchromatic-euchromatic rearrangement, in which a gene, formed by a group of nemameres is breaked (arrow, chromosome a); this resulted in inactivation of two nemameres in a/b (dotted).

After the discovery of the effect of chromocenter heterochromatin on variegation of Drosophila by Schultz, several other observers have tried to determine the extent to which the adjacently relocated zones have been changed. Schultz suggested that euchromatic bands relocated adjacently to chromocenter heterochromatin should have a greater amount of nucleic acid and it has been tried to demonstrate such a difference in salivary chromosomes of mottled strains. This postulate rests on the assumption that the heterochromatin of the salivary glands has a greater amount of nucleic acid; in reality, however, heavy euchromatic bands have an amount of nucleic acid so great or greater than the chromocenter and even in Chironomus the apparently greater amount of nucleoproteins of the heterochromatin located next to the centromere may perhaps be accounted for by a state of greater condensation or "density of structure" (Serra and Queiroz Lopes 1944, Serra 1947). It is not yet proved that heterochromatin has a greater amount of ribonucleic acid, though it is possible that such be the case (Brachet 1938).

According to the known facts, what is to be expected in heterochromatized regions of Drosophila salivaries previously euchromatic, is the demonstration of a meshwork structure when acetic acid is used as a fixative (aceto carmine or orcein); or, alternatively, an overcharge of nucleoproteins which [makes] difficult a resolution into bands, if good fixatives of the nucleoproteins, for instance acetic alcohol 3:1, and a Feulgen test are employed. Cole and Sutton (1941) have not found any appreciable difference in the nucleic acid content between euchromatic bands translocated to near the chromocenter and those which remained en place. However, these authors refer to the difficulty of observing the translocated bands due to their assuming an aspect more or less similar to the chromocenter, that is an heterochromatization. Prokofyeva-Belgovskaya (1947) working with the sc8, wm5, rst5 and wm4 variegated stocks of Dros. melanogaster describes a variable heterochromatization, of the meshwork type, connected with a variegated phenotype, in the euchromatic bands translocated to the chromocenter. For instance, the comparison of the heterochromatization showed by the 3C2 salivary sub-section gave (Table 4 of Prokofyeva-Belgovskaya): heterochromatization of wm4 26.67% of the cells, of wm5 81.25%; mottling, dark mottled in wm4 and light mottled in wm5. As a lighter mottling corresponds to a greater inactivation of the w+ gene, a greater percentage of heterochromatization must be found in it, as really it was. In the series of several sc8 combinations a similar correlation between heterochromatization and variegation has also been found. Schultz (1938, cit. Sutton 1940) has also observed an heterochromatization in certain cases. On the contrary, Sutton (1940) says that only rarely euchromatic bands relocated to near the chromocenter seem to be affected by their new location; however, as a selection against slides with loosely organised bands was performed and the chief effect to be observed was a darkening of the bands, it seems that these observations do not contradict conclusively those of Prokofyeva-Belgovskaya. In order to demonstrate or to disprove an heterochromatization it is necessary to have in mind the characteristic phenotype of the salivary heterochromatin and to make a statistical analysis, since heterochromatization is variable from cell to cell.

The general conclusion from all the observations is that the heterochromatization shown in salivaries parallels the variegation and therefore the mottling, though its genetic consequences are seen only in the affected organ or body part; heterochromatization is indeed a characteristic of the whole individual, not of a particular organ. An investigation of other tissues besides the salivary glands must give interesting results. It is to be expected that in cells where the heterochromatin is not of the meshwork type, an overcharge of nucleoproteins is to be found in the relocated regions.

We turn now to the other aspect of the interactions between euchromatin and heterochromatin. When small fragments of chromocenter heterochromatin are inserted in euchromatic regions, it is to be expected that they assume a less typical meshwork aspect, and conversely, it is only the immediately adjacent euchromatic bands that may become a little heterochromatized. In this case the influence of the euchromatin is greater than that of the heterochromatin. This has been observed to be the case of a weak mottling wm4, which shows a lower grade of heterochromatization of 3C and an inverse degree of heterochromatization of the 20B-F region adjacently relocated (Prokofyeva-Belgovskaya 1947).

Another interesting point is the influence of genotypic and ambiental factors upon the variegation. The rest of the genotype, and particularly the amount of heterochromatin present in the nucleus, exert an action upon the degree of heterochromatization which parallels the effects of the same factors on variegation. It is known (Gowen and Gay, 1933) that the presence of an extra Y, or a Y fragment, causes a lower degree of mottling and the phenotype deviated more to the wild type; parallel to this, it was found that the presence of an extra Y fragment lowers the percentage of heterochromatization in sc8 (Prokofyeva-Belgovskaya 1947).

The age of the flies is another individual factor which has an action upon the variegation, older parents giving rise to an F1 with greater heterochromatization (Prokofyeva-Belgovskaya 1947). This is a very interesting result demonstrating that the aspect of the chromosomes may be influenced by the conditions of the gametes themselves (apparently both of the spermatozoon and the ovulum, the data do not allowing a distinction to be made) as a kind of preformation. As the state of heterochromatization may influence the behaviour of the genes located in certain chromosome regions, this means that certain characters, and particularly mottling and other position effects of one generation, may be influenced by the conditions prevailing in a former generation. The importance of this conclusion and the implications it may have for all Genetics merit that a thorough investigation of individual factors be made on several position effects, concomitantly with the aspect of the chromosomes and the characters shown by the next generations. This should somewhat fill up the gap between lamarckian and mendelian principles and contribute to an understanding of lasting modifications of the Duermodifikationen type. Final conclusions upon these points must await more data.

Of the ambiental factors, it is known that heterochromatization of the salivaries depends upon the temperature: in a case it was found that at 25° heterochromatization is at a minimum [maximum?], while temperatures of 14° and 30° caused a decrease to about half that prevailing at 25° (Prokofyeva-Belgovskaya 1947). Temperature exerts also an effect upon mottling but a correlation with the state of heterochromatization rests to be demonstrated.

The interpretation of these genotypic, individual and ambiental factors seems not difficult in view of what was said above. In principle, all that primarily is capable of influencing the synthesis of matricial nucleoproteins and secondarily of modifying the relations between chromosomes and nucleoplasm should also be capable of influencing the variegation. A greater amount of heterochromatin in the nucleus than the normal probably causes a lower state of heterochromatization at the level of the chromosomes, by competition for a limited supply of the common forming blocks, nucleotides and amino acids, which must be furnished by an unchanged, or almost so, cytoplasm (and also nucleoplasm) quantity. The influence of the individual factors, genotype and age of the parents, is also understandable on this same basis of a cytoplasmic influence. Temperature must influence the synthesis of nucleoproteins and of their building blocks, as other vital syntheses: there is an optimum under the form of a more or less extensive plateau and a steep declivity for both higher and lower temperatures.


The effects of the euchromatic-heterochromatic (eu-het) rearrangements seem to be well explained by the heterochromatization-euchromatization theory, which reasonably fits in the facts known to-date. The question arises, however, if all the other position effects or the effects of rearrangements can be explained by the same theory. In what concerns the mimic effects, which correspond to + alleles being inactivated due to their new neighborhood, it seems quite probable that the euchromatic-euchromatic (eu-eu) rearrangements must have the same explanation as the eu-het ones, except in what respects variegation, which is absent or very rare in eu-eu rearrangements. It is characteristic of eu-eu rearrangements that the position effect usually expands only to over 2 or 3 salivary bands. This and the lack of variegation signify that there is not an extension of the matricial materials, at least in a scale similar to that which characterizes eu-het effects. A simple and apparently plausible interpretation would be that in eu-eu rearrangements the position effect is a direct consequence of the new contacts between non-adjacent regions causing inactivation of the neighboring genes. The difficulty with this explanation, however, is that it does not suffice to postulate an inactivation, but rather it is also necessary to explain how this inactivation could have occurred.

In eu-eu rearrangements with a break near an intercalary heterochromatic block, as for instance probably are the less chromatic bulbs in the salivaries of Drosophila, it would be possible to admit that the inactivation is caused by an extension of this heterochromatization along the chromonema, over a short distance of some bands, and indeed such cases very probably realize in practice. However, the explanation of the eu-eu rearrangements proper, not adjacent to intercalary heterochromatin, must be different: if the conclusions exposed above on the correspondence between factors, chromomeres and bands are accepted and a primary unitary action in phenogenesis may correspond in mean to about 3-5 salivary bands, it follows that any separation of the two "halves" of a gene by a rearrangement could bring about an inactivation—Fig. 3. That is, position effect in these cases would result simply from the fact of the gene having been broken and the pieces separated by a more or less long region of the chromonema. If the relocation took place adjacent to a piece of intercalary heterochromatin, a small extension of the position effect would be possible but ordinarily no variegation should occur due to the fact of the intercalary heterochromatin being less variable in its nucleoprotein charge (as judged from the relative constancy of its structure in salivaries) when compared with the chromocenter.

This hypothesis immediately explains the fact that the location of the second break in a rearrangement may have some influence upon the chief effect. This has been proved in the case of the scutes (Raffel and Muller 1940) for approximately isogenic lines. Another interpretation, namely that the second break more or less affects adjacently located genes which act like modifyers of the principal position effect, is also possible and is preferred by Raffel and Muller. The facts are not inconsistent, also, with admitting that the breakage of the gene causes an almost equal effect, while the details depend upon the exact breakage point within the gene. If, for instance, the break separated bands 1.2 from, or 1.2.3 from 4.5.6, minor modifications could be due to the inactivation more or less extense of the two relocated parts, for instance two bands could be more intensely inactivated than three, i.e. 1.2 more than 1.2.3. The negative results of attempts to obtain by X-rays irradiation a reversion to the original phenotype produced by reversing the rearrangement to the original chromosome sequence in 50,000 male offspring of suitable crosses (Raffel and Muller) though not definitively settling the question, are rather in favour of scute being not an uni-particulate gene.

This explantation of eu-eu rearrangements perfectly fits in with the mean size of 3-5 bands, which we have deduced independently from considerations of the number of genes as compared to that of the bands (see above). If our interpretation of the eu-eu rearrangement effects is correct, it should be expected that the extension of these effects would be also 3-5 bands, as a rule 3 or less, and this is really the extension of these effects observed in practice. This concordance adds much weight to our interpretation.

Another interesting point of this explanation is the possibility of typical eu-eu rearrangements passing from the very limited position effect, not extending more than about 3-5 bands, to a more extensive effect. If several breaks occur adjacently and the rearrangement leads to repeated inactivations along the chromonema, distant not more than about 10 bands, then a "spreading" of the inactivation would be possible and a transition to the eu-het rearrangements should result. These position effects should differ from typical eu-het ones only by their lacking variegation or showing it only in a slight degree. At the measure that inactivation extends over the chromonema, the charge of nucleoproteins could also be maintained during interkinesis and an extension to the sides, accompanied by variegation, would result.

In this view, heterochromatization is a secondary change resulting from a primary inactivation in relatively long pieces of the chromonema. The important relations between heterochromatization, inactivation and mutation will be discussed again below, when attempting to integrate position effects, mutation and eu-heterochromatin changes in an harmonic explanation (see Chapter 6. Gene and chromonema evolution). It must be remarked that also in eu-eu rearrangements can an effect of the rest of the genotype and the presence of extra heterochromatin, as also of individual and ambiental factors, be expected to modify somewhat the phenotype, and especially if the rearrangement is a transition to the eu-het ones. The fact that it has been claimed that temperature alterations seem to have no influence upon ordinary position effects does not contradict this conclusion, as the influence probably is small and detailed investigations are necessary to demonstrate it.


Mimic position effects may be explained by assuming an inactivation of + alleles located near the breaks, which simulate the effect of a mutant. If it is accepted that pure position effects, without equivalent gene mutations, are a reality and not simply a consequence of the respective gene mutation being "not yet" known—an assumption which in Drosophila seems very remote for such intensely studied cases as Bar—these effects must have another explanation than an inactivation of certain loci. Cases of pure position effects are Bar, Hairy wing, Abruptex and Confluens, very likely Star-asteroid and probably also several other cases in Drosophila as for instance Hairless, Henna, Moire, Beaded, Lob, etc. All position effects not linked with deletions and generally dominant, found at loci relatively far from similar mutants, must belong to this category. We think that in future a growing number of position effects will be ranged in the pure type. The majority of lethals, especially the dominants but also a great number of recessives, must also belong to the group of pure position effects of the mimic or the pure types.

Bar, Hair wing, Abruptex and Confluens and Star-asteroid are known to be small tandem duplications or repeats of the direct or reverse types. Lewis (1945) has exposed the view that repeats may provide an explanation of position effects. Although it is not possible to admit that all position effects are of this nature, as is shown by the cases with a Dubinin effect and with variegation, the repeats seem to be an important source of position effects of the pure type. The simplest explanation for the effect of repeats is that they provoke an alteration in the phenotype due to the unbalance caused by dosage of the genes contained in the duplicated piece. In the case of Bar this interpretation has been demonstrated not to hold true and it may be confidently assumed that in other cases an analysis will show that dosage plays no role or only a minor one in causing the outstanding effects of the small duplications. To judge from the number of structures which in the salivaries of Drosophila show the appearance of direct and reverse repeats (Bridges 1935, 1938) this type of position effects must be frequent and repeats would be one of the major evolutionary factors of genetic material (see also Chapter 6).

Deficiencies probably may also cause effects of the pure type, though this is not known with certitude. This could be demonstrated by comparing the effects of def/+ and def/def with def/dupl, dupl/+ and dupl/dupl for the same chromosome piece; if the deficiency was not covered by a corresponding duplication and this latter alone showed no effect or only an effect of another kind, it would be inferred that the deficiency, in addition to the well known mimic effect similar to an extreme gene mutant, caused also a pure position effect. Frequently, deficiencies are heterozygous lethal but, if small, may be heterozygous and hemizygous viable while homozygous lethal. In this latter case of the lethal effect being shown by the homo- and not by the hemizygote, as for instance in several M(1) cases, it may be that a position effect is also involved (other explanations of the influence of the Y, however, may not be excluded before experiments on this point are performed).

The Minutes are also a kind of mutants in many cases demonstrated to be small deficiencies; outstanding is the fact that the Minute effect seems to be non-specific, being the same for several loci scattered over the four chromosomes of Drosophila melanogaster. While the homozygote is lethal, the chief effect in the heterozygous condition is an increase in the developmental time and a lower viability, accompanied by shorter and finer bristles and several other secondary efects on the eyes, wings and body. Some of the M factors are deficiencies of very little pieces without any known mutant locus, while in others a close scrutiny of the salivary chromosomes has failed to reveal any abnormalities. The homozygous lethal effect could be accounted for by assuming that in many loci there are genes synthesizing products, probably of an enzymatic nature, whose lack causes death, as happens in the case of the Neurospora biochemical mutants if certain specific products are not supplied in the culture medium. This interpretation, however, does not explain all the facts, not only the case of some Minutes viable in hemizygous condition, as also the fact that in many a loci no mutant is known, and conversely, that in several cases the small piece deleted corresponds to genes that produce visible, non-lethal, effects and whose lack would hardly account for the lethality. Very probably in these cases there is a complicating pure position effect. Therefore, besides a general "dosage" effect, like for instance that of Haplo-IV, the Minutes may also present position effects.

The case of the numerous Notches is also similar. These hemizygous lethals correspond, according to Demerec et al. (1942) to a deficiency of the 3C7 band of the X chromosome and the effect is not known as a corresponding gene mutation. It was supposed that in this band were localized the genes facet and split; it seems, however, that fa is itself a minute deficiency at 3C7 (Oliver, cot. Goldschmidt 1944). In this way, changes at the 3C7 band cause two separate effects, and possibly a third spl) will be found to be involved also in changes of this thick chromatic band, which may very well have a composite structure. Probably some or all of these effects are pure position effects. It is remarkable that Abruptex and Confluens also are position effects of this same facet region.

The consideration of all these facts on pure position effects suggests the following explanation of them. Pure position effects are the consequence of chromonema components being assembled in a new order, or forming a new sequence, which corresponds to an incipient new gene. Supposing that the rearrangement brings about the junction of bands a.b.c of the gene a.b.c.d.e with bands m.n of the gene m.n.p.q, it is possible that by inactivation of both a.b.c and m.n or of only of one of the groups, the phenotype of the corresponding gene mutants of the inactivation type appear; this is a mimic position effect. If, on the other hand, the assembly of the bands brought about by the rearrangement is different, for instance a.b.c.a.b.c by a direct tandem repeat, or m.d.e.q by an internal deficiency and translocation then it could happen that these new assemblies were capable of giving rise to a new effect during phenogenesis, not hitherto found in the genetic make-up of the species. That is, the species possesses no + allele in relation to these new assemblies of genetic materials, and a pure position effect would be the result.

Pure position effects are, therefore, "genes in the making", essays of creation of new hereditary units capable of self-perpetuation and of modifying the properties of the species. They should be found in relation to repeats and, in a lesser proportion, to small deficiencies, while inversions and translocations, though in some cases giving rise to pure effects, would chiefly contribute to the appearing of the mimic type effects.

If this interpretation is correct, pure position effects of the plus or repeat type should in Drosophila melanogaster be found associated with doublets in the salivary chromosomes, such as, among others, 2B (heterochromatic?), 3C, 4D, F1.2, 8B (hetero- and euchromatic), 11A and C, 12E, 15C, 16A, 18A, 19E, etc. in the X (see the standard salivary maps of Bridges). Effects of the minus or deficiency thpe should be found especially in lighter, though not heterochromatic, regions, for instance 1B10-4, 2C-D, 3A4-9, 4C and D, 6C, 8C, 11E, etc. and in general at the middle of each sub-division, since heavy chromatic bands have been chosen for delimiting sub-divisions. An important corolary of this explanation is that the architecture of the salivary chromosomes, the very aspect, chromaticity and sequence of the bands, is thus linked with the genetic properties of the hereditary materials located in them and a fine causal analysis of this sructure is relevant for the interpretation of the organism. This analysis must be of a statistical oredr and preferently made with a photometric apparatus.

Summarising this discussion on position effect, it may be said that a reasonable interpretation of the known facts must consider two types which at the limits merge into one another: mimic or simulating known mutant genes of the same region; and pure, as a rule different from known gene mutants. The former are yet of two kinds, with possibilities of transitions between them: eu-het (euchromatin-heterochromatin) rearrangements and eu-eu (euchromatin-euchromatin) ones. Typically, the former are linked with variegation, the second not. Both achieve an inactivation of genes, but the eu-het effects may extend over a greater chromonema length, while the eu-eu ones interest only the immediately adjacent genes.

Two causes of inactivation of + alleles in mimic effects are admitted: (1) In typical eu-het effects there is an extension of matricial nucleoproteins, sometimes reaching a considerable distance from the chromocenter, and this ordinarily is linked with variegation. (2) In eu-eu effects the inactivation, though also possibly brought about by local heterochromatization, is chiefly caused by the genes, of a composite nature, being separated into two parts which now are incapable of working as the + allele, and so give a mutant effect of the inactivation type. Transitions between the eu-het and the eu-eu effects and also in their respective causes are found: (a) in the eu-het effects if the heterochromatin is only a small intercalary piece; and (b) in the eu-eu effects if several rearrangements occurred recently (with respect to the time taken by the evolutionary history of the species) in the same chromosome region, which may give rise to extensive inactivation and to heterochromatization. Heterochromatization is thus linked to position effect and by means of this, to gene mutation. Pure position effects are postulated to be caused by rearrangements which bring about a sequence of genetic materials capable of forming a new hereditary unit or gene. This is especially the case of small duplications and to a much lesser extent of small deficiencies, inversions and translocations.

The effect of the forming of a new gene unit is a new element in phenogenesis which has no equivalent + allele and now can mutate. A pure position effect thus results, which is no more than another name of an incipient new gene. Repeats are especially apt to cause these position effects, as they only mildly increase the genetic materials and so make possible gradual modifications in the phenogenetic power of certain chromonema regions.