Controlling Elements and the Gene BARBARA MCCLINTOCK Department of Genetics, Carnegie Institution of Washington, Cold Spring Harbor, New York In a recent brief review (McClintock, 1956), a description was given of types of elements carried in t,he maize chromosomes that serve to control gene action and to induce, at the site of the gene, heritable modifications affecting this action. These elements were initially discovered because t,hey do not remain at one position in t,he chromosome complement. They can appear at new locat,ions and disappear from previously determined locations. The presence of one such element at or near the locus of a known gene may affect t,he action of this gene. In so doing, it need not alter the action potentials of the genie substances at the locus. Therefore, these elements were called controlling elements. It was also shown that controlling element,s fall int'o groups, the members of each operating as an integrated system in the control of gene action. In this report, some aspects of controlling ele- ments mill be considered that could not be dis- cussed in the above-mentioned review. This will necessitate mention of some of the well known gene loci in maize, and symbols for them will be used in the discussion. So that a readv reference may be available, the pertinent i&formation about each of these loci is given in the following list. Chromosome 1: P, pericarp and cob, color. A large number of alleles are known? in- cluding P"", which gives variegated peri- carp and cob color. Chromosome 3 : Al , anthocyanin pigment pro- duced in kernel and plant; al, standard re- cessive allele, no anthocyanin in kernel or plant. Shs , normal development of endosperm tis- sues; sht , recessive allele, shrunken endo- sperm. Very closely linked with Al ; less than a quarter of a crossover unit distal to it. Chromosome 5: Pr, purple aleurone color; pr, recessive allele, red aleurone color. Located in long arm of chromosome 5. AZ , anthocyanin pigment developed in kernel and plant'; at, standard recessive allele, no anthocyanin developed in kernel or plant. Located in short arm of chromosome 5. Gives approximately 28 per cent recombina- tion with Pr. Chromosome G: Y, yellow starch in endosperm; v, recessive allele, white starch. Chromosome 9: I, dominant inhibitor of aleu- rone color in kernel. C, allele of 1, pigment produced in aleurone layer of kernel; c, recessive allele, no pig- ment in aleurone layer. Shl , normal development of endosperm of kernel ; shl , recessive allele, shrunken endo- sperm. Located approximately four cross- over units proximal to I. Bz, purple anthocyanin pigment in plant and aleurone layer of kernel; "62, standard recessive allele, bronze color in both plant and kernel. Located approximately two cross- over units proximal to Shl . WZ, amylose starch produced in pollen and endosperm, stains blue with solutions of I-KI; wx, recessive allele, starch is amylo- pectin, stains red-brown with I-K1 solutions. Located approximately 15 crossover units proximal to Bz. DISTINCTIONS BETWEEX CONTROLLING ELEMENTS AND GENE ELEMENTS In maize, as in other organisms, a change at a particular locus in a chromosome IS made evident by modification of a particular phenot,ype. In- dependently occurring alterations (mutations) at the same locus give rise to change in expression of this one particular phenotype, be it recognized by change in a particular enzymatic reaction or by means of some less well defined but. identifiable modification of phenotypic expression. Because it is possible to predict which phenotypic charac- ter will be altered after modification at a particu- lar known locus in a chromosome, it is inferred that some component is present there whose mode of action may be recognized within certain limits. These components appear to reside at fixed po- sitions in the standard chromosome complement of maize, and they will be referred to in this dis- cussion as the genes. Modification of action of a known gene component can result from insertion of a controlling element at or near the locus of the gene; and, in general, the types of change in phenotypic expression induced by its presence there are those that could be anticipated from previously acquired knowledge of mutant ex- pressions that have resulted from modifications at this locus. Controlling elements, on the other hand, need not occupy fixed positions in the chromosome complement, and detection of the presence of one such element depends upon char- acteristics it exhibits that are independent of its position. It is realized that our present knowledge of the gene does not allow formulation of a definition 197 of it based on structural organization, dimension, or primary type of activitv. However, all the so-called gene loci with w&h \ve will be con- cerned are recognized because a change at the locus effects a change of some component that normally appears in the cytoplasmic region of the cell and therefore the alterat,ion at, the gene locus is reflected in this region. All t,he controlling elements so far identified, on the other hand, may have their area of activity confined within the nucleus itself, for they are known to serve as modifiers, suppressors, or inhibitors of gene ac- tion as well as mutators. They beha\-e as if t,hey were modulators of the genome. Each controlling element or system of interacting elements has its own mode of modulation, and this is expressed in an individualistic manner that is quite inde- pendent of the recognized t,ype of acbion of the gene which it may be modulating. Our present' knowledge would suggest that gene elements and controlling elements represent t,wo different classes of primary components of t,he chromo- some and t,hat a close relationship exists between them. The largest. amount of evidence concerning the mode of behavior of controlling elements has been derived from study of the element called Activator (rlc), which by itself may control gene action at the locus where it resides, and also of another controlling element that responds t,o it. When this latter element is inserted at, a known gene locus, changes in gene action may occur either immediately after its insertion or subsequent to it. Both the insertion of this element at the gene locus and the subsequent modifications in gene action it induces depend on the presence of the AC element somewhere in t,he chromosome com- plement, In the absence of AC no changes affect- ing gene action occur, and st~ability of gene ex- pression will be exhibited as long as it is absent. Return of AC to the nucleus t,hrough appropriate crosses will again effect activation of t,he second element of the system, and this will be expressed in a series of mutation-type changes at the locus of the gene where this element, resides. The second element, of the system outlined above was originally given thra designation Dis- sociation (Ds) because, ill the presence of SC, breaks appeared to be formed at the locus where it, resided. It" was later determined that the ap- parent "breaks" were producaecl becaausr at this locus a dicentric chromatid and a cbnrresponding acentric chromatid were formed. The acentric chromatid, composed of the segment of the chro- mosome from Ds to the end of the arm, was elimi- nated from the nucleus during a mitotic anaphase whenever such an event, occurred. These dicen- tric-acentric chromatids were formed both in somatic and sporogenous cells, and the time dur- ing the development of a tissue when this took place was found to be a function of the dose of AC: t,he higher the dose, the later the time of these occurrences. In the absence of i4c, however, no dicentric-acentric chromatids were produced nor was there any evidence that would suggest the presence of this Ds element at the locus. Return of ilc again initiated these dicentric-acentric chromatid format,ions. It is evident,, therefore, that. test,s of the presence or absence of AC in a plant may be made by crossing it. with one that is homozygous for Ds but does not have AC. These are the so-called r-lc tester stocks, and descrip- tions of their usefulness for detecting the presence of AC have been given elsewhere (RlcClintock, 1951, 1953; Barclay and Brink, 1954j. If these dc tester stocks carry recessive alleles of other known genetic markers, and if the plant carrying 9c is heterozygous for them, it is possible to determine the location of ilc and changes in location that may occur. By this means, various different posi- tions of AC have been detected. Evidence of its transposition from one known location to another has also been obt,ained. In the presence of AC, the Ds element, under- goes transposit'ion, and its insertion at various locations were det,ected. Sometimes LXs was in- serted at or close to a known gene locus, and the `effects of it's presence on gene action were thereby discovered. In some cases both dicentric chroma- tid formations and changes in gene action were noted to occur at this gene locus, but, only when AC was also present in the complement; and the time of t'he occurrences in the development of a tissue reflect'ed the dose of the AC element' that was present in the nucleus. In examining these cases, it was soon learned that the Ds element it'self could undergo modifications that altered its mode of response to -4~. Some of them resulted in a reduced frequency of occurrence of dicentric chromatid formations, often correlated with an increased frequency of occurrence of change in gene action; and these latter jvere unaccotipanied by gross change in chromosome morphology. Still other modifications of this I1s element re- sulted in almost complete elimination of dicentric chromat,id formations, although mutations con- tinued t,o occur at the locus of the gene where the 11s element resided. Thus, in such cases, the designation of "Dissociation" (Dsj for the ele- ment responsible for these mutations no longer appeared t,o be applicable. Severtheless, this original designation has been retained because in t,he early st,udies it was possible to follow sequen- tially the changes in the Ds element that altered its type of response to =Ic: from a high rate of dicentric chromatid formation to a low rate---and also the reverse, from a 101~ rate to a high rate. Therefore, the designation ~`11s" will be .applied to any element, at a known gene locus that re- sponds to AC in the followi-illg manner: in the ab- sence of AC, it undergoes IIO alterations that affect gene action; but in the presence of ilc such alterations occur, and the time of their occurrence during the development of a tissue, and cells of COSTIWLLISG ELEMESTS ASD THE GESE 199 the &sue in which they occur, are a function of dc, particularly of its dose. The designabion "Da" for an element responding to AC should not he construed to mean that this element will produce dicentric chromatids wherever it may be located. It may produce none. Nevertheless, its responses to ;1c, as described above, are readily detected. It is also probable that there are different kinds of "Ds" elements, but all of them respond to AC in this quite predictable manner. It is clear that both the 9c and the Ds ele- ments retain their characteristic modes of ex- pression when located at various different posi- tions in the chromosome complement, and that methods of detecting their presence in these dif- ferent locations have been developed. The study of *-ic and of the integrated Ds-AC two-element system has been of considerable help in investi- gating other unrelated control systems. This ap- plies part,icularly to the well known Dt (Dotted)-ai system originally discovered by Rhoades (1936, 1938, 1941, 1945). The standard recessive, al , is very stable in the absence of Dt. In it,s presence, however, mutations occur. to the higher alleles of .41 or to stable recessives that no longer mutate in the presence of Dt. The pattern of response to Dt is quite predictable. Dots of deep anthocyanin pigment' appear in a nonpigmented background in the kernel, and streaks of anthocyanin appear in a nonpigmented background in the plant. The number of dots (mutations) that appear in the kernel is an expression of the dose of Dt: the higher the dose, t.he more frequent the mutations. The Dt element was located by Rhoades in the short arm of chromosome 9. Subsequently, Nuffer (1955) discovered the presence of Dt in two South American strains of maize. In one st,rain, it was located in chromosome 6 and in the other strain it was located in chromosome 7. The response of the standard ai allele to each of these newly de- tected Dt elements is the same as that expressed when the original Dt element, discovered by Rhoades, is present. Recently, Nuffer (personal communication) obtained evidence of transposi- tion of the Dt element located in chromosome 7 to a new locat,ion in t,he chromosome complement. Thus, in this respect, also, Dt resembles AC; both may undergo transposition without loss of iden- tity. The response of the standard al allele to Dt is similar in many essential respects to that of Ds to iic when the former is present at the locus of a known gene. It could be inferred, then, that a controlling element,, responding to Dt, is also present at the standard al locus and that it is this element which is responsible for the observed mutations and for their pattern of appearance in the plant and kernel tissues. Evidence of this is now available from studies aimed at analyzing t,he composition of the -41 locus made by Laughnan (1952, 1955) and also by Nuffer (personal com- munication). When plants carrying an dl allele in one chromosome 3 and the standard al allele in the homologue are crossed by plants homozy- gous for the latter recessive, some kernels exhibit- ing a mut,ant phenotype appear on the result,ing ear. These express a lower level of intensity of pigmentat,ion than that given by the .41 allele. It has been shown that the majority of these mutants arise as a consequence of crossing over within a compound -41 locus. When plants grown from these pale-colored kernels were crossed by plants carrying Dt, it was discovered that in about 5 to 10 per cent of them the mutant expression was unstable. Mutations to higher alleles of .4r occurred, and the pattern of response (dots of deep pigmentat,ion) was the same as that given by the standard al allele. In these cases, however, the mutat,ions were registered in a pale-colored background rather than in a colorless one. Plants that were homozygous for the dl allele also gave rise to pale mutants. It could he shown that many of these arose from crossing over within the com- pound -41 locus. Sane of them, however, were unstable in the presence of Dt. From this it might he inferred that an element responding to Dt is not present at this -41 locus. Because the pale mutants derived from crossing over in the hetero- zygote wit,h al do carry this element, it may be inferred that those crossovers that occur within a restricted region of the compound locus will introduce it. In the presence of Dt, it will respond in this new organization of the locus in t,he very same manner that it was responding before the crossover occurred. On the self-pollinated ear of a plant that was -4llUl , Dt/Dt in constitution, a single kernel appeared that showed a striking change in pattern of mutation (Kuffer, 1951). -4 very large number of pigmented dots appeared in a colorless back- ground, and t,he intensity of pigmentation in these dots varied from light to very dark. A plant was grown from this kernel, and an investigation of the nature of the change responsible for this altered mutation pat'tern was commenced. It proved to be a new allele of al that responds to the presence of Dt by giving this strikingly differ- ent pattern of mutation: a high frequency of oc- currence of mutation in plant and kernel in the presence of Dt, some of which gives rise to large sectors of mutant tissue and others to small areas. In the absence of Dt, however, no muta- tions occur and the mutant behaves as a stable recessive. By appropriate tests Nuffer (personal communicat,ion) n-as able to learn that the con- trolling element responsible for the dotted pat- tern of mutation and that responsible for this new pattern occupy different sites in the com- pound locus and that' they can be separated by crossing over. It is now evident that the basic mechanism of control of gene action in the Dt-al system is essentially the same as that in the Ds-iic system. Interactions between the members of these two systems do not occur, however, for 200 BARBARA McCLISTOCIi Dt will not substitute for AC in control of Ds (Nuffer, unpublished) nor will AC substitute for Dt in control of gene action at al (McClin- tack, 1953). Recognition of other types of controlling ele- ments is made possible by procedures that are essentially similar to those outlined above. This applies to the Suppressor-mutator system as- sociat,ed with control of gene action at a modified A1 locus (u~~-~, McClintock, 1955, 1956), to a similar system operating to control gene action at AL (McClintock, unpublished), to several sys- tems investigated by Dollinger (1955 and unpub.) , and to still other systems now under investiga- tion in several different laboratories. The presence of controlling elements in the maize chromosome complement is now well established. It is for the fut,ure to determine the extent to which such element,s, and systems of interacting elements, operate in the over-all control of activity of the genome during development. The ever-increasing recognition of these elements and their modes of action suggest that t'hey may perform a major function in t#his respect, and an understanding of the manner by which it is accomplished may be gained from examination of the modes of behavior of particular elements or systems of interrelated elements that are non- known. DETECTIOK OF TRANSPOSITIONS OF CONTROLLING ELEMENTS As stated earlier, the presence of elements, inde- pendent of the genes but controlling their action, was initially detected because they undergo trans- position from one location to another within the chromosome complement, and the effects they produce when inserted at known gene loci are thereby made evident. Therefore, a description of some of the methods that have been used to detect such transpositions is of considerable im- portance for an appreciation of the nature of the behavior of controlling elements. A number of different methods have been used and they fall into two general categories, selective and non- selective. The nonselective method is more labori- ous than the selective method but its use may be required in the-initial study of a controlling ele- ment. For example, in a cross of a plant having a single ,4c element, whose location is unknown, to plants having no AC, half the progeny can be expected to carry an AC element as a result of meiotic segregations in the parent plant. If the individuals in this progeny are tested for the presence of AC by the method described earlier, the expected ratio of presence and absence of AC will usually be found. However, an occasional plant may be present in the progeny that has two ilc elements instead of the expected one. Again, if genetic markers have been introduced into the cross, and if the AC element has shown no linkage with them in the parent plant, the progeny may include an occasional plant which has one AC, as expected, but in which the ,4c element now shon-s linkage with one of the markers (for an example, see Table 1, McClintock, 1951). If the progeny of this plant, in turn, is tested for the location of the ,4c element, the majority of dc-carrying indi- viduals n-ill show the same linkages of AC with the given markers as exhihited by the parent plant. An individual may be found, however, in which this linkage is not expressed, and no linkage with the given markers n-ill be exhibited in its progeny. In this manner, successive changes in location of the Bc element may be detected, but the method is quite unselecti;~e and it requires tests of a number of individuals in successive generations. Once the location of a controlling element has been determined, however, selective methods of detecting subsequent changes in its location may be applied, and these will be con- sidered shortly. A number of different selective methods of de- tecting transposition of controlling elements have been applied, and some of them take advantage of the particular mode of action of the controlling element under investigation. For example, the dose effects produced by the .4c element may be useful in this respect, for it is known that the higher the dose the later the time of occurrence of modifications affecting rlc itself or affecting Ds or Ds-type elements lvherever the latter may be located. Changes in location of AC usually occur rather late in t,he development of the somatic and sporogenous tissues, but oc,casionally they occur in a cell early in plant, develcpment and t.his can result in the appearance of sectors in which t,he AC constitutions differ. If such a sector enters the ear, the altered -4c constitution in its cells is made apparent by the distinctive pheno- types of the kernels that. develop Avi-ithin it after pollen from an AC tester stock has been placed on the silks of the ear. They indicate either that AC is absent in the cells that. formed the sector or that it is increased in number. By selecting kernels from such a sector and by making test crosses with the plants derived from them, it is possible to verify the change in =Ic constitution that has occurred in a somatic cell of the parent plant. Some sectors are twinned, in that, the AC element appears to be absent from one sector and increased in number in the twin; and verification of this is readily obtained from examination of the plants derived from the kernels in each component of the twinned sector. The dose effects produced by AC have also allowed detection of some of -the changes in AC locat,ion that occur late in the development of sporogenous cells and consequently are exhibited only in individual kernels on an ear. Their useful- ness in this respect was described earlier (?tIcClin- tack, 1951) in connection with tests devised to detect such changes in location of AC that occur in plants carrying this element at allelic positions in a pair of homologous chromosomes. They also COSTROLLISG ELEMESTS AA-D THE GENE 201 allow detection of gametes carrying two AC ele- ments, produced by plants having only one. When such a plant is used in a test, cross, the functioning of a gamete having two AC elements gives rise to a kernel that exhibits a much delayed time of occurrence of the dicentric-acentric formations at Ds in chromosome 9 or of mutation at those gene loci where the Ds-AC system is operating. T'erifi- cation of the presence of more than one =Ic ele- ment in these kernels may be obtained by testing the =Ic constitution in the plants derived from them, provided, of course, that the change in --lc constitution occurred in a cell of the parent plant before gamete formation and thus provided for an endosperm and embryo that were alike in .&z constitution. During studies of transposition of Ds and also of mutation occurring at those gene loci in which the IAs-AC system of control of this operates, it has been noted that change in location of AC often accompanies the event that affects the Ds element of the system. Such coincidences are so numerous that, selection of individual kernels exhibiting modification of the Ds element is also useful as a method of selection for changes in location of -1c. One of the most effective methods of selection for transposition of AC utilizes those cases in which the =Ic element resides at t)he locus of the gene \vhose action it is directly controlling: that is, the case of AC (called Modulator and sym- bolized as Xp) at the P locus in chromosome 1, described by Brink and his collaborators (Brink and Silan, 1952; Brink, 19%; Barclay and Brink, 1955; Fradkin and Brink, 1956), and that of AC at the bronze locus in chromosome `9 (AIcClin- tack, 19.56). Mutations at these loci are associated with events occurring to and instigated by the -Jc element itself, and many of them are accom- panicd by removal of .Jc from the affected gene locus and its insertion elsewhere. Therefore, if those kernels on an ear that exhibit' germinal mutations are selected, and the plants grown from them are tested for AC, it will be found that in a number of these plants AC no longer resides at the mutant locus. However, AC may still be present in the chromosome complement but lo- cated elsewhere. Tests of this t'ype were made by Brink and his collaborators with regard to the P locus and by me with regard to the bronze locus. As an example of the t,ype of result obtained from such tests, those conducted by me with the bronze locus n-ill be given here. The phenotypic expression produced when AC is present at the bronze locus resembles that given by t,he standard recessive, bz. This latter recessive is complet,ely stable in the presence of &4c, but mutations occur at the bronze locus where AC resides. Some of these give the dominant Bz ex- pression, and the majority of such mutants are thereafter stable in the presence of SC. A larger fraction of the mutants express the recesslye, bz, and this expression is thereafter stable in the pres- ence of AC. Some mutations, however, give rise to unstable dominants or to other types of change that will be considered later. Here me mill consider only those changes that give rise to mut,ants t,hat are stable in the presence of AC. Kernels exhibiting mutant phenotypes were selected from ears pro- duced by plants carrying only one AC, and it was present at the unstable bronze locus in one chro- mosome 9. The homologous chromosome 9 in these plants carried the standard, stable recessive, bz. The short arm of each chromosome 9 carried other distinguishing genetic markers, to make it possible to ascertain that the observed mutations had occurred at the bronze locus where iic resides. The ears from which the muta&carrying kernels were selected were produced by these plants when they were crossed by plants having no =Ic and homozygous for the recessive alleles of the selected genetic markers. On these ears, those kernels that were homozygous for the standard recessive, bz, were completely bronze; no Bz spots appeared in them. On the other hand, the majority of those kernels that received the chromosome carrying the bronze locus with AC showed a number of deep purple (Bz) spots in a bronze (bz) background. h few kernels having this locus, however, showed an altered phenotypic expression and some of these, in turn, were either totally Bz or totally bz. Be- cause of the constitution of these kernels with respect to the other genet,ic markers carried in chromosome 9, it was possible to refer the changed expression in each case to an event that had oc- curred, in a cell of the heterozygous parent plant, at the bronze locus where AC resides. Some of these kernels were selected from the ears and plants were grown from them. These plants, in turn, were tested for presence or absence of AC, for location of AC if present, and for stability of the mutant expression. Among 16 Bz mutants selected, 14 proved to be stable. In six of the plants derived from these 14 Bz kernels, no AC was present. In five plants, one AC was present, but its position was altered. In four of these five plants, it was no longer linked with genetic markers carried in the short arm of chromosome 9, and in the fifth plant it was very closely linked with WX. In the remaining three plants, AC was present and its position was close to the locus of Bz but probably a short distance to the right of it. Thus, in at least 11 of these 14 cases of muta- tion to stable Bz, the association of mutation with removal of AC from the bronze locus could be established with certainty, and in five of these cases the event could be related to transposition of AC to a new location in the chromosome com- plement. It is probable t'hat transposition of AC to a new location was also associated with the mutation-inducing event in the six cases where AC was absent in the plant derived from a Bz kernel. If this event occurred in a sporogenous cell before meiosis, segregation of the chromosome carrying the AC element in its ne\\- location at the folloming meiotic divisions could have resulted in the production of a gamete in Tyhich the Bz mu- t,ant was present but AC WLS absent. Twenty-four independent, cases of mutation to a stable recessive \vere also examined. In nine of them, no Bc Iv-as present in the gamete that car- ried the stable recessive. In five cases, one AC was present but it sho\ved no linkage Gth markers in the short arm of chromosome 9. In nine cases, one AC \vas present and shelved linkage lvith these markers, In two of these nine cases, AC \vas lo- cated close to Wx, and in one case it \vas located very close to Sh. In the remaining six of these nine cases, its exact location \vas not. determined; it Iyas linked with WX and shelved from 20 to 30 per cent recombination \yith it. In the remaining case of the 24 examined, Wo AC elements \vere present, one located close to bz, and the other showing no linkage lvith markers in the short arm of chromosome 9. Thus, again, in t,his test, muta- tion could be associated \vith a change occurring to the AC element at the bronze locus; for its removal from this locus \vas established \vith cer- tainty in 17 of the 24 cases, and in eight of them the transposition of AC t,o a ne\v location could be determined. It is obvious from the accounts given above that by selection of mutants one can also select for transposit,ions of AC, and that this met,hod is highly efficient. Some\vhat similar results lvere ob- tained in test,s of Ds in the case of cm - 1 (Ds at the C locus). The majority of mutations to C lrere found to be associated \vith removal of I1.s from the C locus, although its insert,ion at, nelv locations could not be det,ected in most, of t,he cases. -\p- propriate genetic markers that lvould have allojved detection of such insertion lvere not. present in the tested plants. Another useful selective method for detecting transpositions of controlling elements takes ad- vantage of crossover techniques. For instance, a case was found in xvhich AC \vas inserted close to but to the left, of 11'~. Various test,s \vere conducted to determine its location by means of crossover techniques. Recombinants appeared in about 10 per cent of the gametes of the plants tested; but this figure did not represent the true crossover value, as the example belo\ \vill illustrate. In one test, the plants having AC had the constitution C SC Wx in one chromosome 9 and C wx and no AC in the homologue; no other -4~ element \vas present in these plants. When they lvere crossed with plants homozygous for I, wx, and Ds (located just to right of WX), and having no AC, the major- ity of the Wx class of kernels on the resulting ear showed variegation for C areas in a colorless back- ground. These kernels carried AC. The C areas were produced by responses of Ds in the I wz-car- rying chromosome to AC, lvhich resulted in dicentric-acentric formations at the locus of Ds. I, carried in the acentric fragment,, \vas eliminated from daughter nuclei in the mitotic division Jvhich follo\ved this event. The majority of the wx class of kernels, on the other hand, jvere totally color- less, since no dicentric-acentric chromatid forma- tions occurred at Ds because no AC it-as present in the nuclei of t'he endosperm. Holyever, a felt of the IVx kernels lvere nonvariegated (no AC) and a fe\v of the wx kernels shelved C areas (AC present). These last t\vo classes of kernels represented the recombinants. When the plants grolvn from the variegated kernels in the WTC class lvere tested for AC location, its position, in about half of t,hem, \vas no longer just to the left of wx, as might be expect,ed from a crossover event. Instead it. oc- cupied a ne\v locat'ion in the chromosome comple- ment. In the other half of these plants, its location If-as that expected from crossing over. In other Tvords, in t.his test, about half of t,he `(recombi- nants" arose not from crossing over but rather from a transposition of AC; and this probably oc- curred before the meiotic divisions and as a con- sequence allo\ved the AC in the ne\v location to be segregat,ed at, meiosis lvith a wx-carrying chromo- some. Other examples illustrating this type of selective method Lvill be considered in the follom- ing discussion of modes of detection of transposi- tions of the controlling element Spm (Suppressor- mutator)' in the Spm-al" - l system of genp cont,rol. A series of studies has been made of trans- position of the Spm element, in the Spm-alm - 1 syst,em of control of gene action at the A1 locus in chromosome 3, and it indicates that t,ransposi- tion of this element occurs lvith a high frequency. In this respect, it resembles the Ds and AC ele- ments, jvhich also may undergo frequent trans- position. The action of the Spm-aim - 1 system \vas outlined in previous reports (McClintock, 1955, 1956), but its origin and mode of operation lvill be revie\ved here. A modification at the stand- ard A1 locus in a sporogenous cell of a plant in a maize culture under investigation resulted in change in action of the genie materials at that, locus. The change \vas discovered in a single kernel on an ear produced by the plant Iv-hen it had been crossed by a plant homozygous for the standard recessive, al . Instead of being fully colored, as expect'ed, this kernel \vas variegated for colored areas in a colorless background. The plant arising from this kernel also exhibited variegation for anthocyanin pigmentation. Subsequent tests indi- cated the mode of control of gene act,ion at this modified A1 locus. The basic mechanism involves tmo controlling elements: one. resides at the modi- fied A1 locus (designated al" - `) and directly con- trols gene action and the types of change in this action t,hat may subsequently occur; and the other is an independently located element, designated Suppressor-mutator (S~WL), lvhich can modify this action in t\vo distinctly different \vays. When the Spm element is present, no anthocyanin pig- ment' is formed. All action of the genie materials COSTROLLIXG ELEMESTS ASD THE GESE 203 at the aim - 1 locus is inhibited by its presence until, in some cells, a mutation occurs at aim - 1 that allows the genie materials to be active, the type of acti\-ity being a reflection of the type of modification produced by the controlling element at aim - `. The esprcssion of the gene induced by this modification is thereafter stable in the pres- ence of Spm and a stable mutation is thereby ef- fected. In the absewe of Spm, on the other hand, the genie materials at the ulm - i locus are capable of some degree of activity and the kernels and plants are uniformly pigmented. This type of gene action at aim - l is quite stable in the absence of Spm, and will be expressed without change in successive plant generations. When, however, Spm is again introduced into the endosperm and zygote nuclei having ulPn - 1 , suppression of gene action is again made evident and the capacity of the Spm element to initiate stable mutation-type changes at, aim- l is also made evident. Unlike ilc, Spm does not show dosage effects, and the number of Spm elements present in a plant is indicated only by progeny tests. -At least' three and probably more independently located Spm elements were present in the initial plant having aim - `, and because of this t.he majority of gametes produced by this plant, and by many of its progeny plants, carried Spm. Its detection was therefore obscured in the initial tests of aim - l. Only after several genera- tions of crosses to tester plants having no Spm was it possible to recognize this element, for in- dividuals were then isolated that had only one or two Spm elements, and the part this element plays in t,he control of gene action and mutation at al" - 1 was then clearly revealed. Clarification of the behavior of this system of control of gene action was thereafter readily accomplished. When plants homozygous for aim - ' and having one Spm element are crossed by plants having no Spm but homozygous for the standard recessive, al , which is st,able in its presence, the ratio of kernel types on the ears produced will usually approximate one uniformly colored (no Spm) to one variegat,ed for deep colored spots in a color- less background (Spm present). If two inde- pendently located Spm elements are present, the ratio of kernel t,ypes approximates one u'niformly colored (no Spm) to three variegated (Spm pres- ent); and if more Spm elements are present, the ratio of kernel types deviates in the expected manner. Again, tests of Spm const,itutions may be made wit.h plants that are homozygous for the standard al allele (and therefore do not directly reveal the presence or absence of Spm) if these plants are crossed by ones homozygous for aim - * but carrying no Spm. The rat,ios of uniformly colored to variegated kernels on the resulting ears reflect the Spm constitutions in the plants. Thus, plants homozygous for al" - 1 and carrying no Spm can serve as tester parents in crosses made for the purpose of determining the presence or absence of Spm in individual plants and its num- bers when present. If the plants being tested are heterozygous for other genetic markers, and if t.he tester stock carries the recessive alleles, linkage of an Spm element with one such marker may be detected after it has been inserted into the chromo- some t,hat carries this marker. By t,his means it was possible to detect linkages of Spm with genetic markers carried in chromosome 5, in chromosome 6. and in chromosome 9. In plants that are ulm - */aim - 1 or aim - `/aI in constitution and that have a single Spm element, large sectors may appear, some of which exhibit t,he pigmented phenotype that is characteristically produced in the absence of Spm. It' can be shown that these sectors arise through loss of the Spm element from a cell early in deve!opment of the plant. The progeny of such a cell may contribute to the development of the ear, producing either all of it or only a part of it,. When the ear is used in a cross with a plant that is homozygous for aim - ' but carries no Spm, in the former case all the kernels on the ear will be uniformly colored (no Spm), and in the latter case all the kernels within a well-defined sector will be uniformly colored (no Spm). If the plants derived from these uniformly colored kernels are again tested for Spm const,itution, its absence in them may be verified. That they carry an ui" - 1 locus capable of responding to Spm may be shown by crossing them to p1ant.s that are homozygous for the stand- ard al allele and have one or more Spm elements. The typical variegated pattern of deeply pig- mented areas in a nonpigmented background will appear in all kernels that have received al" - l from one parent and Spm from the other. In order t'o detect some of the changes in Spm constitution t,hat may occur early in development of a plant, test's were conducted to determine its constitution in the cells that gave rise to an ear on the main stalk of the plant and also in those that produced an ear on one or more of its tillers (side branches). Tests of two ears per plant were obt,ained from 101 plants. In !15 of them, the number of Spm elements was the same in the cells that produced each ear (63 with one Spm; 26 with two Spm; 6 with three Spm). In six plants, t,he Spm constitution was not t,he same in the cells that gave rise to each ear (one case of one Spm in one ear and no Spm in the other; three cases of one Spm in one ear and two Spm in the second; two cases of one Spm in one ear, the second ear having a sector with no Spm). From twelve other plants, tests of three ears per plant were obtained; and correspondence in number of Spm elements was evident in each of the three ears of eleven of them (6 with one Spm; 4 with two Spm; 1 with three Spm). In one plant, the cells that gave rise to two ears carried one Spm element but the cells that gave rise to the third ear had two Spm. Tests were also conducted to determine Spm constitutions in progeny of plants in which one, two, or three Spm elements were known to be present. Those conducted with 249 individuals in the progeny of plants having one Spm will illus- trate the t,ype of result obtained. The parent plants carrying Spm had been crossed by plants that. were homozygous for aim - 1 but had no Spm. A ratio on the result,ing ears of one uniformly pale- colored kernel (no Spm present) to one that showed spots of deep color in a colorless back- ground (Spm present) indicated the presence of one Spm element in the cells that gave rise to the ears. Variegated kernels were selected from these ears, and the plant,s grown from them were again crossed by plants homozygous for ulm - ' but hav- ing no Spm. From this test it was learned that, one Spm element was present in the cells that gave rise to the ear in 215 of these plants. In 20 plants two Spm elements were present, and in six plants three Spm elements were present. In the remaining eight, plants, no Spm was present in any part of the plant. It. appears from these tests that maq of the modifications affecting Spm constitution occur in individual cells relatively late in develop- ment and may occur even in the gametophytic cells of the plants. In order to examine more precisely these changes in Spm constitut,ion, tests were made of t,he progeny produced by plants having one Spm element at, a known location in t,he chromosome complement. Spm m?y be located at' various posi- tions, and several different posit'ions in chromo- some 5, chromosome G, and chromosome 9 have been identified. Tests of the progeny of plants having Spm, at these different locations were con- ducted. Several examples will illustrate the kinds of information such tests can give. One plant that carried a single Spm element was Y/y in constitution. The test) cross made with it indicated that this Spm element was linked with y. On the ear produced by the test cross, one of the t,wo recombinant classes of kernels was Y and exhibited the variegated phenotype (Spm pres- ent). Plants were grown from some of the kernels in this recombinant class, and these, in turn, were tested for Spm const,itution and location. The TABLE 1. TEST CROSS IXDICATISG LIXKAGE OF Spm WITH y- IS CHRO?.IOSOME 6 0 a,m-lShJa:sh?; I' Spur/!/ X ~3 alm-1sh,/alsh2 ; y/y; so Spm I Phenotypes of Kernels Pale Colorless aleu- aleurone (h-0 Spm) Totals Sh, class.. 20 ~ 113 I 30 0 1 271 shz class 10 65 111 ~ ( 591 8 72 291 Totals.. .I 30 1 178 1 170 38 i ii / 72 565 results obtained from one such plant are given in Table 1. This plant, aim - ' Sh2,~alsht , I;`y in constitution, was used as a female parent in a cross with a plant whose constitution was aim - 1 sh,.J ash2 , y/y and which had no Spm element. Link- age of Spm with II was clearly expressed in those kernels carrying al'?' - I in both the Sh, and the .skg classes. -1s expected, kernels homozygous for the standard al allele were completely colorless, for this allele is stable in the presence of Spm. Plants were grow-n from 22 of the variegated kernels in the Sh? IT class, and these in turn were crossed by plants homozygous for aim - I, Xh2 , and y, and having no Spm. This test cross was made in order to determine whether or not Spm would continue to show linkage with I7 and, if so, the number of individuals that would show it. In 19 of these 22 plants, one Spm clement was found tci be present, and it was linked with I7 in each of them. The percentages of recombinant. classes of kernels were similar on the ears pro- duced by all of these plants. ;\mong a total of `7569 kernels produced by these ears, there n-cre 3847 uniformly pale-colored kernels (no Spm), of which G8.3 were J' and 31&l g. A%mong the 371i variegated kernels (Spm present) on these ears, 3033 were J7 and 681 were y. In addition, there were five deeply colored kernels, which may be attributed to germinal mutation at al" - I, for a fen- such mutations are to be expected. The re- combinants were 18.0 per cent of the total, and this percentage compares well with that of the parent ear (see Table 1) \vhirh was 10.3. In two of the 22 tested plants, two Spm elements were present and one of them was linked with I' (39 pale colored, Y: 118 pale colored, !I: 315 varie- gated, Y: 181 variegated, y/!. The remaining plant had one Spm, but it showed no linkage I\-ith I7 (117 pale colored, l': 112 pale colored, y: 101 variegated, I-: 100 variegated, y; and in addition 2 deeply colored kernels attributable to germinal mutation). X much more extended series of tests of the type just outlined was conducted with progenies of plants carrving Spm in chromosome 0 but at another locatron in the chromosome, and these tests were extended into the fourth generation. A description would require more space thall is justified here; it need only be stated that, in gen- eral, the results obtained were similar to those described above. In the interest of further clarifi- cation of the behavior of Spm. however, we should present an example of this kind of test made jvith plants carrying the Spm element in a different chromosome of the complement. In the following case, it was located in chromosome 5. The silks of one ear of a plant that was aim - I Shzialsh? , Prjpr in constitution and carried Spm, received pollen from a plant that was homozygous for al , shy , and pr and had no Spm (cross 1). Another ear of this plant was used in a cross with a plant of similar constitution except that it was homozygous for Pr (cross 2). The kernel tvpes ap- pearing on the ear prodwed by cross 1 iildicated linkage of Spm with Pr. Plants were then grown from some of the variegated kernels in the SS'h:! , Pr class on each of these ears, and these plants, in turn, were tested for Spm constitution. Exam- ples of the results of tests of selreral of these plants are given in Table 2 (plant 668JD-1 and some of its progeny from woss 1, plant 6683F-2 and some of its progeny from woss 2, and also progeny of plant 668X-2 from cross 2). Plant 6684D-1 proved to be al"' - 1 Sh! al.&:! , Prjpr in constitu- tion, and it had one Spm that was located in the chromosome 5 carrying Pr. Its linkage with Pr was evident in all the test crosses (rows 1 to 3 under A4 of Table 2). Plant (%85F-3 n-as also Pr:`pr in constitution, hut the Spm element was linked with pr (B, Table 2). The location of Spm in the pr-carrying chromosome may be at,tributed to a crossover 111 the parent plant carrying Spm, which introduced it into a pr-carrving chromosome. Most of the variegated plants il; culture 6685 (derived from cross 2) were Pr/Pr in constitution. The Spm element in most of them could be ex- pected to be located iu one of the two Pr-carrying chromosomes. The results of tests of one plant of the culture, fi685G-2, which was ulm - 1 Shs,,`u&? , Pri Pr in constitution, were as follows. This plant was crossed by one that was homozygous for aim - I, Sh? , and also for the recessive, pr, and had no Spm. The kernel types on the ear produced by this cross (132 pale-colored kernels: 137 kernels that had deep-colored spots in a colorless back- ground) indicated that plant 6685G-2 had one Spm. All the kernels, however, were Pr in pheno- type. Twenty plants were grown from the varie- gated kernels on this ear, and each mas'tested for Spm constitution and for the linkage of Spm with Pr. Spm was found to be present in 17 of the 20 plants but absent in three of them. Fifteen of these li plants carried a single Spm element; in 1-l of them it was linked with Pr (rows 1 and 2 ucder C of Table 2)) and in one it gave no evidence of linkage with Pr (row 3, C, Table 2). The re- maining two of the 17 Spm-carrying plants had two Spm elements, one of which was linked with Pr (rows 4 and 5, C, Table 2). The kernel types produced by the crosses shown in A of Table 2 indicated that Spm in plant 6684D-1 was located in the chromosome 5 that carried Pr. In order to determine whether or not this linkage would be expressed in the following generation, plants derived from the Pr class of variegated kernels produced by the crosses en- tered in lines 2 and 3 of A of Table 2 were exam- ined for Spm constitution and location. hmong the eleven tested plants in the progeny of the cross given in line 3, one did not have Spm. Each of the remaining ten plants had one Spm, and its linkage with Pr was clearly expressed in nine of them. The phenotypes of the kernels on the ears produced by these plants, after the given test cross, are entered in row 1 under D of Table 2. The Npm element in one plant did not, show link- age with Pr (row 2, D, Table 2). In the second test, all ten of the examined plants derived from the variegated kernels in the Pr class on the ear produced by the cross given in line 2 of -5, Table 2, carried Spm. In nine of them, one Spm was present, and in eight of these it Ii-as obviously linked with Pr (row 3, D, Table 2); in the ninth plant no linkage with Pr was observed (row 4, D, Table 2). In the remaining plant, three Spm ele- ments were present, as the ratio of kernel types entered in row 5, D, Table 2 will indicate. In plant 6685F-3, which was Pr;pr in constitu- tion, S'pm was found to be linked with pr (B, Table 2). Most probably, it was introduced into the pr-carrying cahromosome as a-consequence of crossing over. The variegated kernels in the Pr class on the ear produced b.y the cross given in B, Table 2, represent recombmants, and plants de- rived from ten of them were examined for Spm constitution and location. Spm was found to be present in nine of these ten plants, and absent in one of them. One Spm was present in eight of the nine plants, but in only four of them was it linked with Pr. The combined ratios of kernel types on t,he ears produced by these four plants are entered in row 1 of E, Table 2. In the remaining four plants having one Spm, no linkage with Pr was exhibted (rows 2 and 3, E, Table 2). In one plant, three or four Spm elements probably were present, as the ratio of kernel t'ypes on two tested ears of this plant indicates (row 4, E, Table 2). It was men- tioned earlier, in connection with transposition of AC, that selection of recombinants is an effective method of detecting transposition of t'his control- ling element. The example given above illustrates the usefulness of this method for detecting trans- positions of Spm. A4nother illustration of the ef- fectiveness of this selection method will be out- lined below. A variegated plant that was ulm - `,/aim - I, WX/WX in constitution was crossed by a plant homozygous for uln - l and for wx but carrying no Spm. On the ear produced as the result of this cross, there mere 365 kernels; 196 of them mere uniformly pale colored (no Spm), and 169 had deep-colored spots in a colorless background (Spm present). It may be concluded that one Spm ele- ment was present in the plant t'hat produced this ear. -4mong t,he pale-colored kernels, 156 were WZ and 40 were wx, whereas the ratio in the varie- gated class of kernels was 29 WZ to 140 wx (A, Table 3). Linkage of Spm with the wx allele car- ried in one chromosome 9 is obvious. Nine plants derived from the variegated kernels in the JVX class (the recombinant class) were examined for Spm and for its location. ,411 of them carried Spm, and in seven of them one element was present. In none of the nine ears obtained from test crosses of these seven plants, however, was there any evidence of linkage of Spm with 777x. The com- TABLE 2. TESTS OF Spttr COSSTITYIY~S .4su LWXTIOX IS THE PROGSSY OF I'r~srs (`.UWTISG Spur IS CHROMOSOME 5 See test for origin of plants in A to E below, for the constitutions of plants entered in B to E, and for test crosses made with them. Phenotypes of Iiernels Colorless with spots of deep color: Shl ~ / (Spm present) I Colorless Sk? Germinal 0 I c? ~ Pr P Pr P' : mutant; Shz ~ Tota's I 8. Types of kernels on ears produced by test crosses of plant 6684D-1 which was nl~t~`Shy/alsh~ , Pr Spw/pr in constitution 6684D-1 aghl/a,she ; pr/pr; SO ! 50 Spm alshs/alshp ; pr/pr; ATo 6684D-1 : l2o 1 2: I 2:: I ,:: 1:: I : ; 1:: Spm a,m-`ShJa,m-`Shz ; 6684D -1 80 / 188 114 46 - I / 1 429 pr/pr; h-o Spm I I I Phenotypes of Kernels Spin Constitution of Plants No. of Plants I Pale Colorless with spots of I (So Sjw) deep color I (Sgm present) Germinal mutant Totals Pr P' PI 91 B. 1 Spm I 1 / 2X / 91 I 67 / 214 i I 1 651 C. 1 Spm 6 1 Spm 8 1 Spm I 1 2 Spm 2 Spm 1 1 1 Spm 1 Spm 1 Spm I 1 Spm 3 spm 9 1 8 1 1 434 i 1646 1798 (PF and pr)* 281 (P? and pr)* 49 1 105 1`28 (Pr and pr)* 1512 1429 114 209 132 - - D. 429 67 i 1174 ) 73 1089 (Pr and pr)* 234 (Pr and pr)* 31 ' 36 1133 71 616 98 167 E. 415 9 4016 341 2 3570 132 0 527 148 1 ) 512 87 1 3-B I I 366 3 77 0 ( 3105 2% 251 1 1957 99 1 1 j -I32 I 176 1 0 j 410 / I 1 Spm 4 149 550 510 : 113 1 : 1323 1 Spm 3, 321 290 248 254 1 I 1114 1 Spm 1 185 , 171 134 152 1 6-13 3 or 4 Spm 1 15 I 14 235 246 ) 2 1 512 * In some crosses, difficulty was encountered in discriminating between Pr and pr kernels because of segregation of another factor that modifies pigment color in pr/pr kernels but only, however, in the pale class. 206 COSTROLLISG ELEMESTS ASD THE GESE 207 bined ratio of kernel types on these ears is given in row 1 under B of Table 3. One plant had two Spm elements, but linkage with TTz was not in- dicated in either of the t\vo tested ears of this plant (line 2, B, Table 3). The ratio of kernel types on two ears produced by the remaining plant' indicated the presence in it of three Spm elements, as shown in row 3, B, Table 3. When such a high number .of Spm elements is present in a plant, evidence of linkage of one of them with a given marker is obscured. Thus it cannot be stated whether or not in this plant one of the elements was linked with Wr. It is clear, however, that there is no evidence among these recombinants of linkage of Spm with Vs. It may be concluded that the Spm element in the parent plant was located very close to WC and that the recombinants ap- pearing on the ear of this plant arose mainly from a premeiotic transposition of the Spm element t,o a new location, which allowed it to be segregated at meiosis with the chromosome carrying the WZ allele. In Table 2, the results of tests of 47 variegated plants are entered. The variegated plant was used as female parent in test crosses of 46 of them. In eight of these 46 plants, two ears per plant were used for such tests, and agreement with regard to Spm number and location was shown by the two ears in seven of the eight plants. In one plant, the kernel types on the ear produced by the main stalk indicated the presence of two Spm elements, one of which was linked with Pr, and these are entered in line 5 under C of Table 2. Tests of the tiller ear of this plant indicated the presence of only one Spm in the cells that gave rise to it, and this Spm was linked with Pr (9 pale coiored, Pr: 84 pale colored, pr: G9 variegated, PT: 11 varie- gated, pr). In two additional plants, three ears were used for the test cross, and agreement with regard to Spm number and location was shown in all three ears produced by each plant,. The nine plants t,hat gave the kernel types entered in B of Table 3 were all used as female parents in the test cross, and in four of them two ears per plant were so tested. Agreement with regard to Spm number was shown by both ears produced by each of the four plants. Before we leave the subject of methods of de- tection of transposition of controlling elements, one ot'her fact should be emphasized. If, in any one test, the controlling element is found to be linked to a known marker, and if other known markers are present in the same chromosome, linkage with them will also be exhibited, and in the expected manner. If genetic markers on other chromosomes are also present, and are followed in the test,s, no linkage with them will be exhibited. In other words, the basic mode of inheritance is like that of other known genetic markers, be- cause the controlling element resides at a particu- lar locus until some event causes it to be trans- posed elsewhere, where its position may again be TABLE 3 i I Phenotypes of Kernels I I Spm Constitu- ~~0.~~ tion of Plants A. Test cross indicating linkage of Sp~l nith wz in chromosome 9 0 alm--l/alm--l; IVx/wz Spm X cf alm-l/all"-l; wz/wz; so Spn1 I 1 ) 156 1 40 1 29 1 140 0 365 B. Tests of Spm constitution in D plants derived from variegated kernels in the W'Z class of A, above. 111 plants lvere used as females in crosses with plants that were alm-`/af-l, wx/wx, and that had no Spm 1 spm ~ 7 2 &n/L 1 3 SpnL 1 identified. The time of occurrence of these changes in location, during development of a tissue, and the frequency of their occurrence, depend on several factors: dose (the AC element and the con- trol it exerts on transpositions of Ds), environ- mental conditions such as temperature and nu- trition (Eyster, 1926; Rhoades, 1941; van Schaik, 1954), the genetic background (Brmk, personal communication), and t,he location of the element in the chromosome complement (see account of nontransposing controlling elements in McClin- tack, 1956). THE STRUCTURE OF COSTROLLIKG ELEMEST~ Controlling elements and gene elements are alike in one important aspect, and this is related to the replication of their structural organization during chromosome reduplication. With respect to both types of elements, the newly constituted chromosome is a replica of the parent chromosome, provided no event occurs to alter the structure of the cont.rolling elements or that of the locus where it may rseide; and in most division cycles such events do not occur. The most effective demon- stration of maintenance of organization of a con- trolling element through successive mitotic cycles is provided by that element of a two-element system which undergoes modification only when the second element of the system is also present, such as Ds in the Ds-AC system, or the element at the A1 locus in the al-Dt system and the ulm - I- Spm system. In the absence of AC, Dt, or Spm, in each of these systems the organization at the locus where the complementary controlling ele- ment resides must be replicated without altera- TABLE i. &.%MPLES OF TYPES OF ALLELES OF Ai PRODS-CEU BY THE GPERATIOS OF THE Ds-AC SYSTEM, THE a,-Dt SYSTEM, ASD THE a,m-l- Spm SYSTEM OF COSTROL OF GENE ACTIOX AT THE A, Loccs IS CHROMOSOME 3 A. The Ds-AC System Example: aim+ Phenotype produced in absence of AC Phenotype produced when 1 AC is present Allele 1 ,411ele 1 Colorless kernel. No anthocyanin Dots of the full Al-type espression in restricted regions of kernel. Few pigment in plant. So mutations. streaks of d, pigment in plant. Fen- germinal mutations to higher alleles of Allele 2 i A, . Some chromosome "breaks" at u,"~-~ locus. Allele 2 Uniformly light pale-colored kernel Numerous spots of various sizes showing deep pigmentation in pale- and lightly pigmented plant. So mu- colored background in kernel. Sectors of deep pigmentation in pale-colored tations. background in plant. Many germinal mutations to higher alleles of A, Few if any "breaks" at arm--3 locus. Allele 3 Allele 3 Uniformly pigmented kernel and Similar to allele 2 above escept that mutant spots in kernels and sectors in plant. Intensit) of pigment much plants appear in background coloration of medium intensity. darker than that given by allele 2. No mutations. B. The al-D1 System Phenotype produced in absence of DL Phenotype produced when 1 Dl is present Allele 1 (standard ai) Allele 1 Colorless kernel. Ko anthocyanin in Few dots of deep Al-type in pigmentation in colorless background in ker- plant. No mutations. nel. Fine streaks of AI-type pigmentation in plant. Few germinal mutations to higher alleles of A1 and to stable recessives. hlost germinal mutants are stable in presence of Dt. Allele 2 Allele 2 Colorless kernel. Ko anthocyanin in Large number of pigmented spots or dots in colorless background in ker- plant. So mutations. nel. These show various grades of intensity of pigmentation from light. to deep. Plant shows many sectors of mutant tissue in nonpigmented back- ground. Many germinal mutations to alleles giving various grades of in- tensity of pigmentation, the m3;jorit.y of which are stable in the presence of Dt; a few, however, are unstable. C. The a?"-`-Spm System Phenotype produced in absence of .Spm Phenotype produced in presence of .S)m Allele 1 Allele 1 Colorless kernel. No anthocyanin pigment in plant. Ko mutations. Allele 2 I Very pale color in kernel. Intense anthocyanin pigmentation in plant. No mutations. Allele 3 Many areas of different sizes showing various low levels of intensity of pigmentation in colorless background in kernel and in nonpigmented back- ground in plant. Many germinal mutations to give alleles producing these low levels of pigment intensity. Mutants so obtained are stable in presence of Spm. Allele 2 Small dots of full Al-type pigmentation in colorless background in kernel. Fine streaks of Al-t.ype pigmentation in nonpigmented background in plant. Very few germinal mutations. These are stable in presence of Spm. Allele 3 Intense pigmentation in both kernel and plant. Xo mutations. Allele -1 Many medium-sized dots showing full Al-t,ype pigmentation in colorless background in kernel, and many fine streaks of Al-type pigmentation in nonpigmented background in plant. Fen- germinal mutations to higher d- leles of A i These are stable in presence of Spm. Allele 4 Pale color in kernel. Intense pigmen- tation in plant. X0 mutations. Many spots and large areas of full Al-type pigmentation in nonpigmented background in kernel, and numerous sectors and streaks of Al-type pigmen- tation in nonpigmented background in plant. Many germinal mutations to I higher alleles of Ai that are stable in presence of Spm. - 208 COSTROLLISG ELElIESTS ASD THE GESE "09 tion in each mitosis in the germ line, and because of this it is maintained through successive plant generat,ions. Only when, by appropriate crosses, the second element of the system is introduced, will alterations occur, and then only in some cells of the plant. Illustrations of the type of evidence t,hat has revealed this are given below. When Ds was first discovered, many dicentric- acentric chromatid-forming events were observed at the locus of Ds when AC was also present in the nuclei of a plant. The frequency of their occurrence was high and they could be observed readily in both the plant and the endosperm tissues. Gam- etes having no AC were produced by such plants, and in the next generation those plants having Ds but no AC showedno dicentric-acentric chroma- tid formations at the locus of Ds, nor any other type of change that would reveal its presence. When, however, *4c was reintroduced in a subse- quent generation by an appropriate cross, the presence of Ds was revealed, because dicentric- acentric chromatids were now formed at the pre- viously determined position of Ds, and the fre- quency of their occurrence.mas the same, wit,h the same dose of AC, as that observed before SC had been removed. Besides dicentric-acentric chroma- tid formation at the locus of Ds, other types of change in the presence of .4c mere also not,ed, and some of them effected transposition of Ds to a new location. Occasionally, another type of modi- fication occurred at the locus of Ds, but again only when AC was also present, and this was recognized by a decided change in the relative frequency of occurrence of the above-mentioned types of re- sponse of Ds to AC. The location of Ds, however, was not altered by the event that, was responsible for this. If, by meiotic segregation, AC was re- moored from the chromosome complement that carried such an altered Ds, the particular modifi- cation responsible for it' was maintained without further change through successive plant genera- tions. This was made evident because return of AC in a later generation elicited from it the very same pattern of response t.hat it had given before AC was removed. Through selections based upon such clear-cut changes in the particular types of response of Ds to AC, it was possible to isolate several different alleles of Ds, and to maintain them unaltered in successive generations in plants that did not have AC. The behavior of each of them, when dc was returned to the nucleus, could be predicted in advance if it,s behavior before t'he removal of AC was known. Such predict'ions are possible only if the particular organization at the locus of Ds in each such case is replicated in each successive mitotic cycle in the germ line. It is difficult' in t,he face of such evidence to avoid the conclusion that t,hese alleles of Ds reflect organiza- tional and t)hus structural differences either of t,he Ds element itself or of other components at the locus where it resides, even though the type and dimension of such differences are not yet knoxvn. In all examined cases where a two-element sys- tem of control of gene action is known to operate, changes have been noted in the type of response of the element at the affected gene locus, and iso- lates showing different types of change have been made. The different isolates so far test.ed have behaved in inheritance as alleles of one another; and the t,ype and pattern of response shown by each, when the second element of the system is present,, is predictable. Table -I was constructed in order to illustrate some of the kinds of differ- ences these isolates may exhibit,. Each of the alleles listed in this table demonstrates the presence of a particular type of organization at the gene locus where the controlling element resides. This or- ganization is reproduced in each successive mitosis, and this is maintained through successive plant generations, provided the complementary element of the system concerned is absent. It might be considered that a controlling ele- ment represents some kind of extrachromosomal substance that can attach itself or impress its influence in some manner at various positions in the chromosome complement and so affect the action of the genie substances at these positions. The modes of operation of controlling elements do not' suggest this, however. Rather, they suggest that controlling elements are integral components of the chromosomes themselves, and that' they have specific activities and modes of accomplish- ing them, much as the genes are presumed to have. Any proposed view of the structure and organiza- tion of controlling elements must consider not only the origin and maintenance of distinct alleles, as described above, and the modes of interaction exhibited by two-element systems, but also the nature of the change responsible for alteration in gene action that appears after eit'her insertion or removal of a controlling element at a known gene locus. Insertion need not effect mere inhibition of gene action, as the analysis of the Sprn-c~~~ - I sys- tem illustrates. Also, removal of the identifiable controlling element is not usually accompanied by exact restoration of t,he type of action shown by the genie substances before the controlling ele- ment appeared there, although sometimes it may give rise to a similar type of action. In many cases there is a clearly expressed difference, ranging from slight) to marked. Thus, the presence of a control- ling element at a gene locus need not effect mere inhibition of gene action, and its removal from the locus need not effect mere release of such in- hibition. Other types of modification occur. Some form of organizabional or structural change in chromosome materials must, occur, both when a controlling element is present at a locus and also as a consequence of its removal. The available evidence suggests that the disappearance of an identifiable controlling element from a known location in a chromosome is associated with it's appearance at a new location. The mechanism of transposition, then, is significant in any considera- 210 BARB.ARA McCLISTOCK tion of the structures concerned, and our knowl- edge of this mechanism, inadequate as it is, should be evaluated. THE ~\IECHAXISM OF THA~Y~POSITIOS Transposition of Ds from its first known posi- tion in the short arm of chromosome 9 to another position wit,hin that arm was detected very early in the study of the Ds-AC system, and attempts were made to obtain information about the man- ner by which transposition is accomplished. Be- sides dicentric-active chromatid formation at the locus of Ds, other types of chromosomal aberra- tion were noted in these early studies; they lvere made evident bv the appearance of translocations, inversions, du$ications, ring chromosomes, and deficiencies, but only when AC was also present in the nucleus. In all such cases it was found that one of the two positions in the chromosome or chromosomes involved in the origin of the rear- rangement was always at the previously identified locus of Ds. It was obvious, therefore, that the Ds element was primarily responsible for them. It' was then decided to determine whether or not transposition of Ds accompanied such events, and for this purpose those cases that produced a dupli- cation of a segment of t,he short arm of chromo- some 9 were examined. They were chosen because the genetic markers located in this arm would allow the most precise analyses to be made. Three such cases were examined in detail,. and from all three it was learned that transpositron of Ds had accompanied the event that accomplished the du- plicat.ion. A description of the origin and consti- tution of the chromosome having the duplication, in two of these three cases, was given in an earlier publication (McClint,ock, 1951) and need not be repeated here. It could be determined from a study of the three cases, however, that the Ds element that was inserted into the new location came from only one of two sister chromatids 9. It, was also learned that the insertion of Ds at the new location was accompanied by its removal from its previous location. In each of the three examined cases, the involvement of sister chromatids in t,he origin of the duplication, the orientation of the duplicated segment, and the location of the two Ds elements suggested that contact of the locus of Ds with another locus in this chromosome pre- ceded the event that produced t.he duplication and the transposition of Ds; and that the trans- position could be associated with the mechanism of the subsequent' chromosome reduplication it- self. It was also learned from these cases that insert,ions of Ds into new locations could take place without eff ect.ing gross chromosomal rear- rangements (see positions of Ds in diagrams given in McClintock, 1951). The allele of Ds that gives rise to the types of chromosomal aberrations mentioned above is one that produces many dicentric-acentric chr'omatid formations at, the locus of Ds itself, but only, of course, when Be is also present in the nucleus. It might be considered that such configurations could arise either from lack of reduplication of t'he components at the locus where Ds resides, or from "stickiness" of these materials. The cases men- tioned above indicate, however, that the Ds ele- ment is reduplicated when chromosomal aberra- tions occur, for the two Ds elements produced by the reduplication process can be accounted for even though one of them occupies a new location. This suggests that the dicentric-acentric chroma- tid formations might arise not from lack of re- duplication at the locus but rather as a conse- quence of t,he reduplication mechanism itself. Some component of this particular allele of Ds may be so ordered that' it will allow a reverse bonding between linearly arranged components during the reduplication process; this, in turn, would lead to the dicentric-acentric chromatid formations that are so clearly expressed in some cells of the somatic and sporogenous tissues of plants having the allele. It should be emphasized, in this discussion of transposition, that the Ds element undergoes al- terations which modify its type of action. These mere mentioned earlier. Some of them give rise to an allele of Ds that no longer produces dicentric chromatids (or only a few of them) or other types of chromosomal rearrangement. Transpositions continue to occur, however. The controlling ele- ments AC and Spm also undergo frequent trans- positions and these are not, usually accompanied by chromosomal translocations. Thus it appears that t,ransposition is not the consequence of "stickiness" at the locus where these elements reside; for, if it were, many cases of chromosomal aberration should have been detected in associa- tion with their transposition. The absence or in- frequence of such cases is conspicuous. One can conclude, then, that transposit,ions of controlling elementas eit'her arise from some yet- unknown mechanism or occur during the chromo- some reduplication process itself and are a con- sequence of it. At present, the latter interpretation is favored, for itwill account for many of the ob- servations of change in number and location of controlling elements. If, in a plant having one such element at a known location, the element is transposed from one of two sister chromatids of one chromosome to a nen- location in one of t,wo sister chromatids of another chromosome, segrega- tion at the following mitotic anaphase could re- sult in several different types of change in consti- tution of the element in the .sister.nuclei. Either one nucleus would have this element at the pre- vious known location and the sister nucleus would have it at a new location, or one nucleus would have no element and the sister nucleus would have tlvo, one at. the previous known location and one at a new location. As a consequence of such segre- gations, cells could be formed having no such ele- ment, or one that was unchanged in its location, (`OSTROLLISG ELEMESTS hSD THE GESE 211 or one at a new location, or two elements, one at the previously know1 location and one at a new location. In the gametes produced by plants having a transposable element, just such types of change in location and number of elements have been found; esamples were given earlier. Changes in constitution of AC occurring in so- matic cells give rise to sectors that eshibit these changes, and in some of them it is clear that the dose of AC has been increased. It is possible to observe t,hat subsequent changes in dc consti- tution may also own, for subsectors showing them may appear within the larger sectors. If sequential transpositions of a controlling ele ment occur in the cells of the germ line in the manner outlined above, gametes carrying more than two element,s may be produced by plants whose zygote nuclei had but one; and examples of this have been found (see examples under D and E of Table 2). If transposition occurs during the chromosome reduplication process, then t,he means by which it is accomplished is of considerable importance. It is conceivable t,hat it is brought about by some mechanism similar to that proposed to account for transduction in bacteria (Demerec, Blomstrand, and Demerec, 19%; Demerec and Demerec, 1956) -a form of exchange or "crossing over" bet,ween the component contributed by the phage and t'hat present in the bacterium, which requires a re- duplication of both components. If such an event accounts for transposition, t'hen reciprocal sub- stitutions of components at the loci concerned should be produced as a consequence. There is evidence to suggest that such substitutions may occur. In the study of AC at the bronze locus, de- scribed earlier, three rases of change in mode of control of mutation at this locus were not,ed. Each was associated with removal of AC from the bronze locus, and in t,wo of the three cases it could be es- tablished with certainty that AC had been trans- posed to a new location. In two of the three cases, a recessive, bronze, phenotype appeared as a con- sequence of the removal of -4~; but' instead of be- ing stable in its presence, as in most such cases, this phenotype was stable only in its absence. In its presence, mut,ations to the higher alleles of Bz occurred, and the type of response was the same as that exhibited in other cases in which the Ds-AC system of control of gene action operates. The mode of control in the third case was similar. In this case, however, removal of AC was associated with partial expression of the genie substance at the bronze locus, for some &-type pigment ap- peared both in the plants and in the kernels hav- ing it. This expression was quite stable in the ab- sence of AC, but in its presence mutations occurred to give alleles that are associated with the ap- pearance of higher or lower levels of intensity of the pigment. The origin of these three cases could be explained if a &-type element (one respond- ing to AC) had been substituted for AC during the transposition process. There is other evidence that should be men- tioned in considering the possibility that substi- tution may accompany transposition, and this is related to the types of mutation that are produced when a known controlling element is removed from a locus. As stated earlier, the mutations so pro- duced need not be alike; in some cases, wide dif- ferences in their mode of expression can be ob- served. It is possible that some of the differences they exhibit are the consequence of substitution of one t,ype of controlling element for another. Only through further investigation, however, moult it be possible to verify this; and except in the three above-mentioned cases evidence is not yet available, although in several other cases there is evidence suggestive of it. POSITIOM IN THE CHROMOSOME COMPLEMEKT `AT WHICH CONTROLLIKG ELEMEP;TS MAY BE INSERTED It is known that, controlling element)s may be inserted at various locat,ions within the chromo- some complement. In order to learn whether these positions are randomly distributed or selectively locat,ed, a large number of independent transposi- Cons of a particular element from a known loca- tion to new locations needs to be determined. Even though many sequential transpositions of the ele- ments Ds, L4cT and Spm have been detected, the evidence obtained from any one of these elements is insutlicient to allow definite conclusions bo be drawn. The evidence does suggest a degree of nonrandomness, which, however, may merely re- flect degrees of viabilit>y following upon insertions at particular positions rather than selectivity of positions at which the elements may be inserted. Such inviabilities were discovered in studies aimed at detecting the positions within the short arm of chromosome 9 into which Ds may enter. This arm carries genetic markers that allow easy detection in kernels of insertion of Ds between any two of them. Plants having Ds at a known location in this arm produce some gametes having Ds at new locations, some of which are also in the same arm. When plants carrying Ds and also dominant alleles of some of the known genet.ic markers in this arm were crossed to plants that were homozygous for the recessive alleles, kernels that arose from func- tioning of a pollen grain in which Ds occupied a new position in the arm mere readily detected. The endosperm and embryo of a number of such kernels were quit'e normal in appearance, but other kernels were abnormal in various ways. Some of them were smaller than normal, others were germless or had defective embryos, and still others exhibited distorted growth in the endo- sperm tissues. Most of these kernels did not ger- minate. Some of the normal-appearing kernels also did not germinate. In other words, dominant lethality was exhibited among a number of kernels 212 BARBSRA McCLISTOCK having Ds at new posit,ions \Tit,hin the short' arm of chromosome 9. The new position of Ds could be verified only in plant,s derived from the kernels that did germinate. The positions occupied by Ds in these plants seemed not to be randomly dis- tributed but to be clustered about certain locations within the arm. It is suspected that, they repre- sented only some of the positions int,o which Ds may enter, and that the insertion of Ds at other positions results in dominant lethalit,y. This in- ference is supported by the results of examinations of cases that exhibited semidominant lethality of the following type. In endosperms with two nor- mal chromosomes 9 not carrying Ds and a chromo- some 9 with Ds at the new locat,ion, growth was so distorted that many kernels failed to mature. Only a few reached mat,urity but all of them were obviously aberrant. in morphology. Plants could be obtained from some of the lat.ter, however. Four independent cases of this type were exam- ined, and in all four the semilethal effect was associated with a modification induced by Ds in a chromosome component. located to the right of Bz. In these plants, the chromosome 9 carrying Ds was quite normal in morphology. The semi- dominant let,halit,y was not due, primarily, to inhibition of gene action, for it is known that endo- sperms that are deficient for all of the short, arm of one chrotnosome 9 develop normally. Some change in gene action other than localized inhibi- tion was induced in these cases. Because dominant' lethality may be expressed when cont,rolling elements are inserted at some positions in the chromosome complement, as de- scribed above, difficulties are encountered in at- tempts t,o determine whether a particular element can enter any site in the chromosome complement or is restricted to certain sites. At present, no definit,e st,at)ement, can be made regarding this. It is known, however, that controlling elements may be inserted at a number of different positions. Those that seem to be preferred, on the basis of present knowledge, may represent only a selected number of possible sites at, which the presence of the element does not induce inviability at some stage in development. INFLUENCE OF COSTROLLISG ELEMEXTS IK MODIFYING GENE ACTIOX It is now known that controlling elements may modify gene action in a number of different ways. They may influence the time of gene action in the development, of a tissue, and also determine the cells in which it will occur. Again, they may in- fluence the tvpe of action, with regard to either degree (quantitative aspects) or kind (qualitative) aspects). They may also act as inhibitors, sup- pressors, and modifiers, as lye11 as inducers of types of change at a gene locus that resemble those often referred to in the past as point, or gene mutation. When a particular controlling element is in- serted at a gene locus, its presence there may be detected by the changes in phenotypic expression that appear. Tests may t'hen be conducted to de- termine the posit,ion in a chromosome at which the responsible changes are orcurring, and thus the locus involved may be identified. Subsequent tests may be made to examine the mode of operation of the particular controlling system concerned with these changes and to determine it,s components. From such procedures it was learned that one system can operate at a number of different gene loci, and that the action at one gene locus may be controlled by different systems; t,he evidence has been presented in previous publications (RIcClin- tack, 1953, 1955, 1956). When a controlling element is inserted at t,he locus of a known gene, a recognizable change in phenot.ypic expression may be observed as an im- mediate consequence, or no immediate change may result. In t'he lat'ter case, the presence of a controlling element at' the locus is detected sub- sequent,ly, for it will init,iate recognizable changes in gene action. The origins of al" - 3 and aim - 4 will illustrate this fact. Both arose from insert,ion of a Us-type element (one responding to AC) at the standard A1 locus in chromosome 3, and the Ds-Ac Mo-element system controls gene action in both cases. In the case of al, - 1, inhibition of gene action probably occurred as an immediate conse- quence of insertion of a Ds element at the standard ill locus, as its presence there was made evident a few cell generations after that event occurred. In the case of aim - 3, on the other hand, insertion of t#he Ds element at t'he 81 locus did not produce an immediate change in gene act.ion. Its presence \vas revealed later, holvever, by altered expressions of the gene substance at the locus, which were recog- nized in some of the progeny of one particular plant of a culture. At, least, a full plant generation intervened between insertion of the Ds element at A1 and recognition of its presence there. Had ,4c been removed from the nucleus shortly after this insertion, the presence of the Ds element at the locus \vould not have been detected, for in the absence of AC the phenot,ypic expression would have remained unaltered. Only by some fort,uitous cross that again introduced =Ic into a plant carry- ing this original state of aim - 3 xvould the pres- ence of Ds have been revealed, for only then n-ould frequent change in gene action occur. Thus, the presence of a controlling element at a particwlar locus is revealed by the types of change that occur under given conditions, and if these conditions do not prevail its presence at the locus may not be recognized. That the presence of a controlling element at a locus need not effect inhibition of gene action is also shown by studies of those cases in which LIC resides at the bronze locus in chromosome 9 and at the 1' locus in chromosome 1. The phenotype produced when these cases were initially recog- nized was that of the recessive, or null, expression. In both cases, mutations occurred, and the part COSTROLLISG ELEAIESTS Ah-D THE GENE 213 that AC plays in these was reviewed earlier. It was found that removal of -AC from the locus concerned was associated in some cases with a rhange in gene action from one that gives the null expression to one that gives a high degree of activity, and also that the mutant so formed was subsequently stable in the presence of .-lc when the latter was located elsewhere. Some of the mutants were not stable, however, and subsequent mutations oc- curred. In the case of AC at, the bronze locus, two of the 16 Bz mutants examined were of this type. From studies of both of them it, was learned that the event that gave rise to the Bz expression did not result in removal of AC from the immediate vicinity of the Bz locus and that =Ic was responsi- ble for the subsequent mutat,ions t'hat occurred. Some of these later mutations resulted in stability of the Bz expression, and in t,he t,hree examined cases it was learned that AC had been removed from the locus of Br. Other changes were detected by the reappearance of the unstable recessive, and in the several esamined cases it was learned t.hat AC was still present at the bronze locus. Still other types of mutant espression were not,ed, but these need not be discussed here. It is desired only to emphasize that the presence of a controlling ele- ment at a gene locus need not effect inhibition of action but may instead condit,ion a mode of con- trol of gene action in subsequent cell and plant' generations, which will follow in a predictable manner. The ability to predict depends, of course, upon the extent of knowledge of the controlling system in operation in any one case. The presence of a particular controlling element at a known gene locus can influence gene expres- sion in different ways, which may range'from com- plete suppression to various degrees of action. Moreover, the types of action may differ not onlv quant'itatively but also qualitatively. There is evi- dence to suggest that in some cases these various types of mutation are reflections of modifications affecting different components of a compound locus, and that each component of the locus is concerned in its own way with development of one particular phenotype. Extensive evidence based on crossover studies of the compound nature of the A1 locus in chromosome 3 has been pre- sented by Laughnan (1919, 1952, 1955a, b) and evidence regarding the R locus in chromosome 10 has been obtained by Stadler and colleagues (Stad- ler and Xuffer, 1953; Stadler and Emmerling, 195-1, 1956). Instability of expression of 81 has appeared rather frequently, and cases of this have been examined by several maize geneticists (Rhoades, 1936, 1938, 1941, 1945; LlcClintock, 1951, 1953, 19%; Xuffer, 1955, 1956; Laughnan, 1956; Peterson, 1956; Richardson, 1956). On the basis of our present knowledge of the origin and behavior of such "unstable loci" it is inferred, when not determined with certainty, that a con- trolling element resides at t'he di locus in each such case. The modifications in gene action it induces there can affect the action of one or another of the known components of this locus, or it may affect all of them simultaneously. Other gene loci, such as that of C in chromosome 9, also appear to be compound. Evidence of this was obtained by study of several distinguishable types of mutation that occur in the case of cm - 2. (For origin of this case and the controlling system involved, see WClintock, 1951.) These are associated with the production of at least t,wo different diffusible substances, both of which are required for pig- ment formation. It was also noted that the dose espression given by the C allele commonly used in genetic studies is related to the limited production of one of these subst,ances by this allele: t,he more C alleles present, the greater t,he amount of this substance formed, and, consequently, the denser the pigmentation. It is conceivable, then, that some of the qualitative differences m expression of mutants of a given locus reflect alterations in ac- tion of different components of a compound locus, and that a controlling element, as the consequence of one event, may affect the action of only one component or of more t,han one; or the modifica- tion induced by any one event may affect the action of all of them. It is non- known that the presence of a control- ling element at a known locus can effect change in gene action not only of the genie components lo- cated close to it, but also of other genie compo- nents located some distance to either side of it. A number of examples of this kind affecting gene action in a particular segment within the short arm of chromosome 9, extending over a region 5 or more crossover units in length have been exam- ined; and these have been reviewed in previous publications (McClintock, 1953,195-l, 1955,1956). Recentlv, Richardson (1956) found a case of "spreading effect" that appeared to be induced by a controlling element located at 8i in chromo- some 3. The nature of the changes responsible for such spread of mutat'ion along the chromosome is of considerable importance for an understanding of the manner by which controlling elements can induce their effects, and those involving the seg- ment of chromosome 9 that includes the loci of I, Sh, and Bz are particularly useful for this pur- pose. The "spreading effect" in these cases is known to be induced by the presence of a Ds ele- ment that is located just to the left of Sh; and it is also known that the mutat,ion-inducing process is not accompanied by change in location of Ds. One might consider that the "spreading effect" is merely the expression of a deficiency for the loci involved, even though most such cases give viable homozygotes, or, barring this, that the or- ganization of the chromosome segment concerned or the structure or organization of components within it is altered in some particular manner by the Ds element. Therefore, tests of some of these cases are being conducted in order to determine whet,her or not crossing over occurs within the 214 BARBARA ;\icCI~ISTOCIi affected segment. It has been learned, in several of these cases, that either crossing over does not occur or its frequency of occurrence is very low. In others, however, crossing over takes place within the affected segment and the frequency of occurrence differs among the several examined cases. Plants derived from reciprocal crossovers are now under investigation in order to determine if mutant expressions of genie components within the affect'ed segment will appear in their progeny, and if so, the types each may give. That separa- tion by crossing over can occur between the in- dividual genie components whose modification is responsible for the compound mutant expression has already been shown by Richardson in the case of the " spreading effect ," mentioned above. Examination of the mode of operation of two- element controlling systems has revealed the breadth of influence of such systems in modifying gene action. The study has concerned not only the number of different genes that such systems control, but also the manner of this control, which is of greater significance. It has been possible to examine the mode of operation of the Ds-AC sys- tem at seven different gene loci (for references, see McClintock, 1953, 1956), and to learn that the system operates in essentially the same way in each case. The mode of control of gene action resides in the system itself. It is the Ds-type ele- ment at the locus of the gene that is directly re- sponsible for control of gene action and for the changes that occur in gene action; and it is the AC element of the system that is responsible for initiating these modifications where the Ds ele- ment resides, and also for the time of their occur- rence. In some of the cases examined, the change in gene action is usually associated with removal of Ds from the gene locus, and stability of mutant expression in the presence of AC is thereby effected (F - l, bzm - l are examples). In other cases, the Ds element is not usually removed when a change in gene action is initiated, and, in the presence of AC, subsequent changes may occur (cm - 2, wzm - l, m - 3 are examples). Any one of the mutants so zoduced, however, will be stable in the absence of AC. Thus, by removing AC from the nucleus, it is possible to isolate a number of alleles that are distinguishable from each other by different modes of gene expression (see Table 4 for examples). Extensive examination of the operation of this Ds-AC system as well as other systems has thus provided a large body of knowledge, from which it is possible to conclude that controlling systems are composed of distinct and well-defined entities in the nucleus, that these are independent of the gene elements as defined earlier, that they need not reside at fixed sites in the chromosome com- plement, that they retain their identities when transposed from one location to another, and that they operate in much the same general manner wherever they may be located. The mode of operat'ion of the Spm-al" - 1 two- element system, which was considered in some detail earlier in this discussion, is impressive be- cause one aspect of control of gene action expressed by the Spm element of the system is suggestive of the mode of operation of suppressors, inhibitors, and modifiers that have been identified in other organisms. In its presence, all action at the aim - l locus is suppressed, but in its absence all but one of the alleles of aim - 1 are active to some degree, and some of them produce kernels and plants that exhibit intense pigmentation (see Table 4). The Spm element undergoes many transpositions n-ith- out suffering loss of identity, for its mode of con- trol of gene action at aim - ' has been found to be the same when variously located. In contrast to AC in the Ds-AC system, increment.s of Spm do not effect modification in time of mutation at al" - I. The same phenotypes appear when either one or more Spm elements are present in the nu- clei of a plant. Another two-element system has been examined in which the mode of control of the independently located element resembles, in certain respects, those of bot,h Spm and AC. The element in this system that is comparable to Bc, Spm, or D1, exhibits a suppressor-mutator type of control of gene action and mut,ation at a modified A2 locus, designated a2" - l, and in a manner that is similar to that of the Spm element of the aim - l system. For example, in its absence, gene action at the azm - l locus in one of the isolat,es (one of the alleles of azm - `) resembles that given by the standard A2 , for the plants and kernels are intensely and uniformly pigmented and no mutations occur. In its presence, however, all gene activity at azm - l is suppressed until the occurrence there of a muta- tion-inducing event that allows pigment to be produced; and the time of occurrence of such events during the development of a tissue depends on the dose of this suppressor-mutator-type ele- ment that is present in t,he nucleus: the higher the dose, the later the time of occurrence. In this respect, it very much resembles AC. Differences in phenot,ype produced by incre- ments of AC or of the Spm-type clement just de- scribed are sharply defined. However, the effects of additions of elements that exhibit this type of dosage expression need not be so contrasting. It is known from other evidence that these effects may be expressed. in a somewhat different manner, and an example will be given here. It involves two elements, one of which resides at the A2 locus in chromosome 5; the other element is independently located. A low dose of the independently located element of this system effects early-occurring mutations at the locus of A2 , from a nonactive allele to one that gives an apparently standard A2-type expression. Added increments of this element effect step-wise delays in time of occur- rence of these mutations, until mere specks of pigment appear in a nonpigmented background in the kernel and only very fine streaks of an- COSTROLLING ELEMEXTS AKD THE GESE 215 thocyanin pigment appear in a nonpigmented background in the plant. Increments of the element above that which gives this speckled pattern result in a striking change in pheno- typic expression. Sow, pigment, is produced in both plant and kernel. Although t'he intensitv of color is low, the pigment is uniformly dis- tributed and no mutattions occur. This change in gene action is nqt the consequence of mutation at the modified At locus, but rather the expression of a high dose of the independently located ele- ment of the system. This may be shown by cross- ing the pale-colored plants to standard a2 tester stocks in which the clement is absent. Doses of it are thereby reduced in some of the progeny, and this reduction is evidenced by the reappearance in kernels and plants of the variegated phenotype: a nonpigmented background in which pigmented spots or areas appear. It has also been determined that' this element undergoes transposition. This may sometimes be exhibited in sectors in the plant or kernel, whichexpress the change in num- ber of the element that occurred in the ancestor cell giving rise t,o t,he sector. It was st,ated earlier that the presence of a con- trolling element at a particular gene locus may not be recognized unless favorable conditions for re- vealing it are also present; and examples were given. The origin of a number of mutants might well be traced to effects produced by controlling elements, but this might be as difficult as it ini- tially appeared to be for the standard al mutant in maize. This mutant responds to Dt by produc- ing higher alleles of A441 , but in the absence of Dt the recessive mutant expression is most stable. The mutant was used in genetic st,udies long be- fore Dt was discovered or its control over the ele- ment at the al locus recognized. Had Dt not been discovered, the at mutant would still be considered an esample of stable gene change, just as many of the ot'her k~m~n mutants are nom considered. In order to identify readily the presence of con- trolling elements and to be able to examine their different modes of behavior, only those cases were selected, originally, that gave clear-cut evidence of their presence by modification of gene action in both somatic and sporogenous cells. These are the so-called " mutable genes" or " mutable loci." It is quite evident,, however, that the standard a1 mu- tant, is as good a memher of the "mutable gene" class as any other that has been described in this or previous reports. How many other known mu- tants, whose behavior appears to be stable in t,he existing experimental cultures, also belong to this class? Can conditions be altered so as to expose the presence of a controlling element? Efforts in t'his direction have not yet been made, except with the standard al mutant itself (McClintock, 195la, b). The method used gave positive results in that case; and this suggests that, positive results might also be obtained if similar experiments were conducted with other known mutants which ap- i;i; to be st,able in the genetic stocks now being C&trolling elements appear to reflect the pres- ence in the nucleus of highly integrated systems operating to control gene action. The modes of operation of the known two-element systems bring into sharp relief one level of this integration. Other levels are nom under investigation, and t,hese are related to t'he effects produced by modifiers of particular systems that have appeared within the cultures under examination or been introduced by strain crosses. For example, in the Spm-carrying cultures, a certain type of modifier arises with rather constant frequencies. In its action it re- sembles, in certain respects, the Spm element it- self, and it may be derived from this element in some manner. It differs from the original Spm, however, in quite recognizable ways. In the ab- sence of Spm, this modifier effects suppression of gene action at a? - I, but only in the aleurone layer of the kernel. The kernels may be totally colorless or they may shorn one or several small dots of deep color. The plants, on the other hand, show the pigmented phenotype that is exhibited in the absence of Spm. In the presence of Spm, the modifier behaves as a recessive, and in test crosses of plants carrying both of tfhem typical segrega- tion ratios for both elements are exhibited. The modifier probably undergoes transposition just as Spm, does, for it has been located at several dif- ferent positions in bhe chromosome complement. It is a controlling element and also it is a compo- nent of this system. Still other types of modifiers of this system have appeared, each expressing a characteristic type of modification of the system, or, in other words, a characteristic type of inte- grative action within the system as a whole. Recognition of the presence and operation of a so-called two-element system, then, represents only recognition of the lowest integrative level of those elements in the chromosome complement that are directly concerned with modification of the genome as a whole. Nendelizing units associated with phenotypic expressions that are similar in many essential respects to those produced by known controlling elements in maize, such as modifiers, suppressors, and some t,ypes of inhibitors, have been described in a number of organisms. Are many of their ef- fects also attributable to the activities of control- ling elements; and what kinds of criteria may be used to discriminate between the proposed two classes of genetic elements, that is, gene elements and controlling elements? Transposability, which made possible the recognition of controlling ele- ments in the chromosome complement of maize, may not serve in all cases as a reliable criterion for discrimination between the two classes of ele- ments, because the frequency of its occurrence may be so low, under certain conditions, that detection may be difficult (McClintock, 1956). Nevertheless, as far as my knowledge goes, little 216 BARBARA McCLIKTOCK if any effort has been made to detect transposition of mutators, modifiers, suppressors, or some types of inhibitors in other organisms, and the degree to which it may occur is not yet known. It would be surprising, indeed, if controlling elements were not found in other organisms, for their prevalence in maize is now well established. REFEREKCES BARCLAY, P. C., and BRINK, R. A., 1954, The relation between Modulator and Activator in maize. Proc. Nat. Acad. Sci. Wash. 40: 1118-1126. BRINK R. A., 1954, Very light variegated pericarp in m&ze. Genetics 39: 724-740. BRINK, R. A., and NILAX, R. A., 1952, The relation be- tween light variegated and medium variegated peri- carp in maize. Genetics 37: 519-544. DEMEREC, M.., BLOOMSTRAIW, I., and DEMEREC, Z. E., 1955, Evidence of complex loci in Salmonella. Proc. Nat. Acad. Sci. Wash. 41: 359-364. DEMEREC, M., and DEMEREC, Z. E., 1956, Analysis of linkage relationships in Salmonella by transduction techniques. Brookhaven Symp. Biol. 8: 75-87. DOLLINGER, E. J., 1955, Analysis of a mutable system in maize. (Abstract.) Genetics 40: 570. EYSTER? W. H., 1926, The effect of environment on variegated patterns in maize pericarp. Genetlcs 11: 372-386. FRADKIN, C. W., and BRINK, R. A., 1956, Effect of Modu- lator on the frequency of endosperm mosaics in maize. Amer. J. Bot. 43: 267-273. LAUGHNAN, J. R., 1948, The action of allelic forms of the gene A in maize. I. Studies of variability, dosage and dominance relations. The divergent character of the series. Genetics 53: 488-517. 1949, The action of allelic forms of the gene A in maize. II. The relation of crossing over to mutation of A*. Proc. Nat. Acad. Sci. Wash. 35: 167-178. 1952, The action of allelic forms of the gene A in maize. IV. On the compound nature of A* and the occur- rence and action of its AD derivatives. Genetics 37: 375-395. 1955, Structural and functional aspects of the A* complexes in maize. I. Evidence for structural and functional variability among complexes of different geographic origins. Proc. Nat. Acad. Sci. Wash. 41: 78-84. 1955, Structural and functional bases for the action of the A alleles in maize. Amer. Nat. 89: 91-103. 1955, Intrachromosomal associat,ion between members of an adjacent serial duplication as a possible basis for the presumed gene mutations from A* complexes. (Abstract.) Genetics 400: 580. 1956, A possible clue to the nature of Dt action. ;\laize Genetics Cooperation News Letter 30: 67-68. MCCLINTOCK, B., 1951, Chromosome organizationand I genie expression. Cold Spring Harbor Symp. Quant. Biol. 16: 1347. 1953, Induction of instability at selected loci in maize. Genetics 38: 579-599. 1953, Rlutation in maize. Carnegie Inst. Wash. Year Book 52: 227-237. 1954. RIutat,ions in maize. Carneeie Inst. Wash. Year Book 63: 25&260. 1955, Controlled mutation in maize. Carnegie Inst. Wash. Year Book 54: 245-255. 1956, Intranuclear systems controlling gene action and mutation. Brookhaven Svmn. Biol. 8: 58-74. NUFFER, M. G., 1951, A highly mutable allele of A1 . Maize Genetics Cooperation News Letter 25: 38-39. 1955, Dosage effect. of multiple Dt loci on mutation of a in the maize endosperm. Science 121: 399400. 1956, Instability of the alpha and beta components of the A1 locus. Maize Genetics Cooperation sews Let- ter SO: 101. PETERSOS, P. A., 1956, An al mutable arising in pgm stocks. Maize Genetics Cooperation News Letter 30: 82. RICHARDSOS,. D. L., 1956, An unstable compound locus, all. Maize Genetics Cooperation News Letter 30: 111-118. RHOADES, M. M., 1936, The effect of varying gene dosage on aleurone color in maize. J. Genet. 54: 347-354. 1938, Effect of the Dt gene on mutability of the al al- lele in maize. Genetics 83: 377-397. 1941, The genetic control of mutability in maize. Cold Spring Harbor Symp. Quant. Biol. 9: 138-144. 1945, On the genetic control of mutability in maize. Proc. Sat. Acad. Sci. Wash. St: 91-95. VAX SCHAIK, T., 1954, Elect of plant vigor on variega- tion of pericarp. Maize Genetics Cooperation News Letter 28: 81-82. STADLER, L. J., and EYMERLIXG, M., 1954, Problems of gene structure. III. Relation of unequal crossing over to the interdeoendence of Kr elements (S) and (P), (Abstract.) Science 119: 585. , 1956, Relation of unequal crossing over to the interde- r;i-y3n7ce of R* elements (P) and (S). Genetlcs 41: STADLER, L: J., and NUFFER, RI. G., 1953, Problems of gene structure. II. Separation of nr elements (S) and (P) by unequal crossing over. Science 117: 471- 472. DISCUSSIOK CATCHESIDE: I w&ld like to suggest t,hat Dr. McClintock has in the transposition of controlling elements the explanat,ion of the phenomenon which has been referred to by several speakers as negative interference. It. would follow that evi- dence for controlling elements in organisms other than maize should be sought in those cases which show apparent negative interference.