Linkage Experiments with Mediums

saaiiaiaiiaen sty se! the American Society for Psychical Research ‘olume 60 — April 1966 — Number 2 Linkage Experiments with Mediums Kartuts Osis! ABSTRACT: Experiments using mediums as subjects were aimed at the problem of evidecvs foe aacetyal ater dette, Intensive explorations were fase to find Ways in which classical mediumistic phenomena, such as messages about deceased 2S aie i a pyeangle Nala modern experimental ilot experiment in which a famous biologist was the supposed communi- In pilot xn and appointment methods were incorporated into a mate didlas, ie, aaa human links was interposed between the mediums and Solent s Me (the living source of information). In order to block the possi- bility of ESP between the medium and the widow, each participant was aware only of the adjacent link in the chain.Some mediums gave some apparently verid- ee EET about the biologist, in spite of the linkage barrier to ESP between iving. In two further experiments with the personality of the deceased Ross, cross crepes wae exore ug the nk din. Rasies rescunbling cian assessment methods were also explored in these two Satenies onion The main ent incorporated the findings of these methodological explo- rations into a acorigvotl balanced-block design experiment. The results were evaluated by an elaborate analysis of variance technique. Only marginal si was shown in some factor interactions. ‘The perspective for future approaches to such research is discussed and found encouraging. cator, INTRODUCTION The question of personal survival after death is an immensely complex one and a great variety of empirical approaches to the problem, and theories about it, have been presented in the literature of psychical research. Some investigators cite as evidence against the survival theory findings which show that personality traits and memory traces are dependent on the body’s neurological structures and neurophysiological processes (7, 8, 13, 18, 20), while others present parapsychological observations of v g validity in efforts to find factual support for the survival hypothesis (2, 3, 6, 13, 14, 17, 19, 22, 34). The types of confirming evidence which seem most impressive concern apparitions collectively perceived (10, 12, 30, 35, 40), memories of supposed previous incarnations (36), spontaneous cases which seem to occur between a living person and a deceased one, exhibiting intent on the part of the dead (27), PK cases osten- sibly showing activity by the deceased (28), and visions seen by zenearon Tepusieg here was ormed while the author was on the 4 of the pPisacrhoboes Laboratory, University, and later during his association with the Parapsychology Foundation, Inc.

92 Journal of the American Society for Psychical Research dying persons (4, 23), The behavior of mediums during trance is also of speci cee in considering the survival hypothesis. Mediums assume the personality characteristics of deceased persons with whom they are sppneky i communication (11) or produce messages which seem to be drawn from their memory ranges (14). These phenomena are especially interesting when the data are gathered in experiments using sophisticated methods rae rad sittings (1, 37, 38, 39, 41) or cross correspondences (16, 24, 31). My own conviction, however, is that research in no one of these areas alone will provide definite answers to the survival question, but that all of them make n contributions to an evaluation of the complex whole, There is no “crucial experiment” which, if the outcome were clear-cut, would decide the survival question once and for all. The problem requires more reliable and better differentiated knowledge in all these areas, and ps even in areas as yet undreamed of. Gardner Murphy has explained the burning need for new data very aptly: Where, then, doI stand? To this the is: what happens when an irresistible force strikes an immovable object ?' To me, the evidence cannot be by-passed, nor on the other hand can conviction be achieved. It is trivial and childish to ask whether I believe fifty-five— forty-five or forty-five—fifty-five. Trained as a fea eee and now in my sixties, I do not actually anticipate ing myself in existence after physical death. If this is the answer that the reader wants, he can have it. But if this means that in a serious philosophical argument I would plead the antisurvival case, the conclusion is erroneous. I linger because I cannot cross the stream, We need far more evidence; we need new ive; perhaps we need more courageous minds (21, p. 273). BACKGROUND OF THE LINKAGE EXPERIMENTS This paper will discuss only one narrowly delineated approach to the survival problem, a series of experiments which concentrated on obtaining from mediums (under conditions to be described later) veridical messages ostensibly drawn from the memory and interest ranges of the deceased. Assuming that it is possible to receive such messages, experimental situations were designed which we hoped would tend to separate information obtained b ESP from EvOE sources from that obtained from the decease It must be emphasized that in this research the question of survival after death is the problem under investigation, i.c., left However, for the sake of simplicity we will use a terminology which implies survival. Terms include the following: Experiments with Mediums 93 1. A medium (M) is a person claiming to produce psi phenomena some regularity and believing that these productions originate m a contact with the dead. In our experimental design, Ms used at two stages: as “appointment mediums” to arrange the situations, and as subjects to carry on the actual tests. 2. An appointment medium (AM) is an M who helps to arrange : test situation by making an “appointment” with the deceased iy a preliminary contact. The AMs were not the actual subjects _ 3. The subject (S) is the person who is experimented upon, Often he is called the “receiver” or “‘percipient” in the literature, ‘mediums were used as subjects in our experiments and they _ herei referred to as Ss. _ 4. The agent (A), often called the “‘sender’’ in the literature, is the who is the source of the information received by the S _ §. The Hving agent (LA) is the living n who is the possible source of the psi information obtained by the S. In part, the LA could himself be the ‘‘stimulus’’ or ESP “target” for the S; in art, his memories about a deceased person could be the source. LA is often called the “‘sitter’’ in the literature. 6. The deceased agent (pAy is the hypothetical surviving per- sonality assumed to be the S’s source of information in certain of mediumistic experiments. The DA is often called the unicator” in the literature. 7. The agent's proxy (AP) is the person who sits in for the LA in a session with an S. The AP knows nothing about the DA and therefore cannot be a source of information about him, In the literature the AP is called the “proxy sitter.” In the present series laboratory assistants were as APs. Experimental Methods T searched the classical research literature on mediumship to discover which experimental methods seemed best able to facilitate the S’s performance; my special interest was in arrangements which could be used in a modern experimental design. Three methods seemed to be promising: 1. The proxy sitting method, which attempts to introduce a barrier to ESP between the S and the living source of information. 2. The appointment method, which was devised by Nea Walker (41) to sensitize the DA to a particular S. 3. The cross correspondence method. These arrangements will be explained in the following pages.

94 Journal of the American Society for Psychical Research Expansion of the Proxy Sitting Method into the Linkage Experiment The psychical research literature contains a good deal of evidence that gifted Ss are able to give facts concerning deceased co about whom they have no normal knowledge (5, 11, 14, 15, 33). It has long been recognized, however, that such facts do not provide strong evidence for survival because the possibility exists that they could have been obtained by ESP between the S and an LA who knew the deceased person. We cannot assume a DA when ro could have been the source f the ificalty by obtained. ier research tried to minimize this difficulty arranging sittings in which the relative or friend of the deceased was not present at the sitting, but represented by a proxy who knew nothing about the desired communicator. Teepettiy between the S and the AP, therefore, could not yield useful information about the deceased ; and since the LA was absent, telepathy between him and the S would presumably be more difficult. Figure 1 shows (a) the usual sitter-present arrangement and (b) the proxy arrangement. « b, L____Sabert (Matin) | _et (Medem)_| a ae Fic. 1. Graphic representation of sitter present and proxy sitting arrangements. The main objection to the otherwise useful proxy method is that the S's might have the power to surmount the single barrier of the AP by discovering from his mind the identity and location of the distant relative who would be a living source of the information about the desired communicator; the S could then draw the information from the memory and surroundings of this relative. To make such an effect more di we decided to use an ex- tension of the proxy method by putting more minds between the S and the living source of information. In such a situation, for example, the LA, a relative of the desired communicator, could ask his acquaintance, Jones, to contact an experimenter, Smith, who turns to an assistant, Brown, who then conducts a sitting with an S. Everyone in this chain knows only the adjacent person. In this case, the S’s ESP would have to surmount a series of human links in order to get the information known to the LA. By manipulating the number of links used (see Figure 2), the researcher can attempt to segregate living sources of information from deceased sources. Two possibilities exist: Experiments with Mediums 95 (1) If it is found that the linkage block cuts the communication line between the S and the living source of information and at the same time does not interfere with the flow of veridical material bout the deceased, then the h esis of communication between the S and the deceased would gain considerably in certainty. ») If, on the other hand, it is found that the information obtained about the deceased is not veridical when the connection between the S and the LA is blocked, then we might assume that all com- aunication in unblocked conditions is based upon ESP between he S and a living source. { of the above interpretations would be valid unless the akoge block satisfied two specifications: (a) that it really de- (Rl te leaded i Beers and (b) that it was “such a nature that it did not interfere with ESP between the S and the DA. : ; _ Once the decision had been made that the linkage idea would provide a substantial basis for our experiments, we searched for ea secondary methods which might facilitate communication between e S and the DA without breaking the blocks which (it was hoped) would tend to eliminate ESP between the S and the LA. In reviewing the various methods employed in mediumistic experiments, I was struck by the method of Nea Walker, assistant to Sir Oliver Lodge. This method consisted of holding a sitting at which a poe homag hy 7a wuarana meer sas ad certain specified future sitting. In his introduction to Miss cer’s book, Through a Stranger’s Hands, Kenneth Richmond describes the method as follows: ee a which is made to de- ceased friends of Miss Walker's, who have professed to cooperate

96 Journal of the American Society for Psychical Research in these experiments ... these “Requests'’ are written down, and also spoken aloud, but privately, sometimes with the cooperation of Miss Walker's two sisters, who are scienti interested in the undertaking. But the method is less that of a private sitting than of an endeavor to put the few identifying details within reach of Miss Wales ag pri who ~ aed to find has communicators requi y istant sitter] and to ensure their presence at an arranged sitting with Mrs. Leonard (41, pp. 14-15). Miss Walker herself gives a more concrete description of her technique in the same volume: My method in attempting to get evidence through Mrs. Leonard (the medium] about a communicator unknown to me, has been to write out a letter to the Group [deceased agents who had been friends of Miss Walker] just as I should to friends on earth, explaining what I want, and telling what little I know. This, of course, is not communi- cated to Mrs. Leonard; everything that she knows about anything is noted in reports of sittings, and she knows nothing of my “requests.” PM A Mes ae Mccall eh Be ie grec in the Prichard case ran as follows: “April 22nd, 1930. ‘This is just to record that I have asked [i ing mentally and aloud when alone] for the brother of a Miss Kate Prichard, to come to my next Leonard sitting on ox Say (Saturday) afternoon. His sister lives at .... She is twenty-five . .. . He was twenty-three, and was killed in Malay ccggclby tar yin Lhe 1929. This is all I know about him. Please try and and bring him.” Such letters are read over by me, usually the last thing at night, and aloud, when I am alone (41, pp. 24-25), Miss Walker’s method seemed to provide an answer to our need. Although it sounds fantastic to say that it is possible to make an appointment with a deceased person as we e a business ap- pointment to meet someone at a certain time and place, the evidence supplied by Miss Walker raised hopes that it might work. In an case, despite wide inquiry, neither my colleagues nor I could think of anything which might be more promising and therefore we veg to build the appointment method into our experimental gn. The Cross Correspondence Method Another method of designing experiments seemed to be useful for our purposes: the use of cross correspondences. In brief, cross correspondences are messages, purporting to come from a DA, which are fragmented or broken into parts. Each part is given to a different S at a different sitting. Usually the parts little sense by themselves and only form a coherent unit when taken Experiments with Mediums 97 her. The cross correspondence idea has been well described urphy (17,21), Roll (29), and Saltmarsh (31), who were rigued some of the classical S.P.R. cross correspondence results (16, 24, 31) possibili of ESP between the living is not completely out in classical cross correspondence experiments, ever: telepathy between an LA and the Ss could have caused results. We theorized that by using a linkage design we could virtually eliminate the possibility that any one of the participating Ss, or the experimenter, could have brought about the results by telepathic impact on the other Ss. We decided to incorporate the classical cross correspondence technique into the appointment- finkage design by using more than one S. We could then check each S’s statements for messages similar or complementary to the ones received by the other Ss. As already stated, the experiments reported here incorporate the t method and the cross correspondence method into linkage design. In general, the plan worked as follows: We st had the appointment medium (AM) make contact with the DA from whom communications were desired. We provided the AM with all the known aids to ESP communication: an interested sitter who was present, psychometric objects owned by the DA, } phs of hag and so forth. When the 4M had supposedly contact with the DA, the AM set up a future appointment between the DA and one or more Ss who were unknown to the AM. The DA was given a date, hour, and minute. He was asked to agree on certain si instance, a nickname or a first-and- second name combination—which would te oka — J It was ible to use the appointment-li esign in two ways: Binet we could use it to pi information from one S about the DA and his life events. Secondly, we could set up appoint- ments between two or more Ss and a single DA and then try to find cross correspondences in the information supposedly geen by the same DA to the various Ss. In our experiments we tried techniques. Testing and Assessment Techniques Mediumistic statements are not easily assessed. Researchers have found that the simple annotation methods of the early experi- ments are inadequate and have replaced them with various ry methods involving quantitative assessment, such as those of Pratt (25), Pratt-Birge (26), and Saltmarsh-Soal (32).

98 Journal of the American Society for Psychical Research None of these methods were found to be suitable for evaluating experiments in which several variables are com and by ity, therefore, I had to explore other possibilities: (1) restric- tion of the stimulus field, (2) tracer effects, and (3) global assessment. Restriction of the stimulus field. In conducting survival experi- ments the investigator can give the S free rein to make statements about anything and connected with the DA, or he can limit the S to obtaining answers to questions concerning very limited aspects of the DA's life. If the investigator chooses the latter course, he can verify the statements concisely and work out appropriate statistics. Numerical data such as birthdays, anni- versaries, ages, etc,, are especially useful because it is ble to calculate the probability of their being correctly stated by chance. As will be seen below, we attempted to restrict the stimulus field through the use of a questionnaire. Tracer effects. Jan Ehrenwald (9) has devised the idea of tracer effects. This pine le was used in some of the explorations de- scribed here. In it, the experimenter sensitizes the telepathic agent, living or dead, to selected words, phrases, or numbers which would be very unlikely to occur in the normal course of the sitting, It is then possible to assess the probability of these tracers occurring in the statements of the S. In our preliminary experiments we used a three-digit number and words randomly selected from Webster's Dictionary as tracers. Global assessment. In contrast to restricted stimulus items and tracer effects, all the S's statements about a DA can be analyzed as a unit. We found that the global assessment method provided data amenable to complex statistical analysis, but our first explo- rations did not lend themselves to anything so ambitious. In the main experiment, however, we used this evaluative technique. Exploratory Approaches Since we wanted to explore the applicability of our three experi- mental principles to research on the survival problem, we decided to try out the general ideas and the technical testing methods in a small pilot series of tests before going on to a more complex experimental design geared to highly structured data collection methods and elaborate statistical analyses. PRELIMINARY LINKAGE EXPERIMENTS Explorations with a Questionnaire We made our first explorations of the linkage idea with the widow of an internationally-known biologist. She acted as an Experiments with Mediums 99 AM and as the LA as well. She and her husband had been a closely- = couple and had cultivated telepathy between each other, informed me. As an example of this, she said that since they on an island which had no shops, whenever she needed certain groceries she would concentrate on the required items and her husband would arrive home from the mainland with the right She also that her husband communicated with after his death through her own and a friend’s automatic ‘It seemed appropriate to try out the linkage idea with this couple. We wanted to sce whether Ss, either individually or in cross correspondences, could produce information about the biologist. On my behalf, three assistants made arrangements with the Ss in this exploration: An assistant in Philadelphia for the peration of three Ss and also participated as an S herself; an in Munich arranged sittings with two Ss in Germany one in Switzerland; and an assistant in New York had a sub- : in England arrange sittings with five Ss and also arranged one directly. In the case of the English arrangements there were two links between myself and five of the six participating Ss. A total of thirteen Ss icipated (see Figure 3). | A cccmartanierad rhoive-cyuesttineito ten’ Stletiovioed - these questions, which concerned the biologist’s life and work, and Fic. 3. Links and participants in the “Biologist” experiment.

100 Yournal of the American Society for Psychical Research had them approved by the widow, presumably with the corrob- oration (through automatic writing) of her husband. Instructions to the Ss and the questions were as follows: Experiment A This experiment is organized on a scale involvi glad to have your cooperation in this first toward Ss is ced piohibien cf baietval afies dent anes Instructions The main idea of the experiment is to get thro’ if you will concentrate on a key sentence and on his photograph. The communicator is accustomed to control automatic writing; pgircat befor yeber tte aphid wey ied tae Sy om ie Toe rescrion Yam pene We 6 OnatenGe OF 1607 What did you use to say to your wife when bringi 3. Nickname of your wife? ™ Be Se! 4. When is your wife’s birthday? sta birthday? was your occupation ? . ep we work mainly . What is sentence you gave? 9. On what were you working at the time you had to leave this wo! 10. Where did you take your last trip in the earthly life? Il. re opeet you once when watching the movie “David Copper- 12. Please give a sentence characteristic of yourself. I knew the correct answers to questions 6, 7, and 8 only. _After the experiment was over, the Ss complained that the highly structured test was “asking too much” and indicated that it seemed to inhibit the flow of imagery. Nevertheless, some in- teresting statements emerged. For example, one S got the correct year of birth of the widow. Another S got this same year, but attributed it to the husband; she also stated the right month of birth for both and missed the day of the month by two for the Li Experiments with Mediums 101 ‘widow, by three for the husband.* Countries and other localities were incorrectly given except by one English S, who correctly placed the biologist as a teacher at Stanford University, an out- standing hit. Another S came close to making a direct hit on the question concerning the communicator’s last work when she eplied that he had been painting a portrait. Actually, he had been painting a still life which prominently featured the sculpture of a head, Although these statements must be considered incon- clusive as evidence concerning the source of information since ‘they occurred mixed with a number of incorrect items, they en- couraged us to explore other avenues. Li Explorations with Tracer Effects and Cross Correspondences The pilot experiment just described taught us that most of the Ss disliked the questionnaire technique; it was therefore abandoned. The statements such as those concerning the birth dates and the Stanford University hit indicated possible success using tracer elements. In order to make the data amenable to statistical eval- uation, the tracers were selected by a random method. The pilot experiment also revealed slight indications of possible cross corres- and strengthened our wish to continue this line of igation. We seized on an opportunity for further experi- mentation when Mrs. A, a Texas businessman's wife who had been engaged in psychic activities most of her life, reported her experiences to the Parapsychology Laboratory at Duke University. Events Leading to Cross Correspondences Mrs. A related the following events: Ross, a businessman friend of hers, had committed suicide in October, 1953, after a heavy financial loss. After his death he appeared in her dreams and visions. These “visits” consisted of conversations. At one point in a dream she challenged his identity, I quote in part from her report: Early in March of 1955, I dreamed that Ross was with me. I said, “J just can’t believe you are really here.” [He replied] “So you want me to it, I su ?” His voice was teasing. “Yes, give me a sign,” I said. “W: the return of your locket do?” Several years before I had left my baby locket while visiting his mother’s home. He asked me if he could keep it for good luck. I think he carried it for a while in his wallet,

102 Journal of the American Society for Psychical Research I told my friend Mrs. B about the dream. A few days later I visited in New Orleans. In a telephone conversation with my husband I asked if I’d received any mail. “You had a letter and pees it,” he said. “It had nothing in it except an old-fashioned Me In answer to my inquiry, Mrs, A’s friend, Mrs. B, wrote to me as follows: This is how I remember it—A called me one morning and said, “Ross was here last night.” I asked if they talked and she said, “He told me someone had my little locket and it would be sent to me.” I interviewed Mrs. B later in New York; her verbal account was consistent with the information given in her letter. Mrs. A’s phone call to Mrs. B occurred several days before the locket was returned. The incident was discussed between them, both before and after the arrival of the locket. Of course, the locket had not come from heaven, but from Canada! Ross’ widow, who had not been in contact in any way with Mrs. A for years, suddenly sent it to her. The important point is that the widow sent it a week or two after Mrs. A’s dream and not at any other time between October, 1953 (when Ross died) and March, 1955. In answer to my query, Mrs. A wrote: “The dream was probably in the first week of March; the necklace probably arrived the middle of the month.” There is a certain vagueness in Mrs. A’s memory of the incident, but it seems likely that the locket was mailed after the dream occurred. In her dream Mrs, A challenged Ross to give proof that he was “really there’’ and he promised to “return” the locket. There was no hidden meaning or obscure symbolism—it was a straight- forward predictive promise followed by fulfillment. If, in view of the timing involved, we accept that there was a causal connection between the promise made and the return of the locket, then several alternative explanations for the incident can be tentatively considered: (1) Precognition on the part of Mrs. A. But this does not satis- factorily account for the fact that Ross’ widow made the sudden decision to return the locket at that particular time, (2) his (nag influence exerted on Ross’ widow either by (a) Mrs. A or by (b) the surviving onality of Ross. In regard to (a), very little is known about the occurrence of telepathy between hostile persons hostility characterized the relationship between Mrs. A and * widow), but it would appear to be rare; as for (b), it involves the difficulty of having to assume survival after death. Thus, agit we may eliminate the precognition hypothesis, we are not able to make a firm choice between the two hypotheses in- Linkage Experiments with Mediums 103 volving telepathic influence, On the surface, however, it looks as though the motivation behind the incident was Ross’ intention to assure Mrs. A of his continuing existence. ‘ In addition to this dramatic instance of ostensible contact with the deceased Ross, Mrs. A rted other psychic experiences connected with him. In 1962 I ed that six months before his death she had a dream foreshadowing it. In answer to my question whether her experiences were dreams or visions, Mrs. A wrote: The dreams were the beginnings of visions, I’d say. For instance, ae ices | Gras ensening that on we the office to get a box that Ross had mailed us—on opening the box I found I wrapped a leaf about my finger and the leaf died. Then the box turned to a casket and Ross was in it. I awakened screaming and it—the dream Ea vetoes cctaieawen Otay | Wee ete co Co te ee ean details of Ross’ death so I couldn’t have asleep. In the promise of the return of the necklace I was asleep. In this account a box containing something precious (orchids) is associated in a dream with a coffin and Ross’ death. At the time of the experiment I did not know of this association (I found out about it only in 1962), but it was one of the leitmotifs in the cross correspondences which were found in the experiment we conducted based on Mrs. A’s supposed contact with Ross. The Cross Correspondence Experiments First experiment: We decided that Mrs. A's contact with the Ross personality seemed definite enough to provide the basis for a second series of pilot tests incorporating our experimental principles. Mrs. A was agreeable so we organized the first session for Feb 12, 1956. In contrast to the earlier exploration (with the widow of the California biologist), in these tests we put emphasis on the Ss’ spontaneous, free responses to the entire Ross personality. However, we used two tracers which we had selected with the help of a random number table: the number 472 and the word “faint.” Mrs. A, acting as an AM, was told to impress Ross with the idea that he should communicate these tracers in addition to anything else he might choose. In addition, Mrs. A gave me a few words which, according to her, Ross had su | as his own tracers. We did not emphasize cross correspon: in this , but ho for them. tthe perth = a part of the instructions to the Ss was as follows: The main idea of this experiment is to get impressions from a certain communicator who wants to communicate through you if

104 Journal of the American Society for Psychical Research you will concentrate on the name Ross at...... o'clock in the after- noon of...... hho s.9/s 934 The communication method preferred by yourself should be used. No questions will be asked [of the communicator], Please try to open up your mind to the impressions coming from that person at that particular time, The contents of the communication will supposedly be limited to simple, even trivial ea communicator might want to “get through” certain words or show a certain thing or might help to draw i etc. He will also try to communicate stvieo digs i meee The three assistants in Philadelphia, New York, and Munich, and the sub-assistant in England who had helped in the earlier work, made arrangements with three Ss for this session; two of ae Peassencnil eg Naa as Ss, bringing the total to five. Two Ss cipated in revious experiment; i and one in Switzerland a Figure 4), iad seni Fic. 4, Links and participants in the first “Ross” experiment (February, 1956), Results: Only two of the five Ss indicated a three-digit number; Ganigeh thee loth. tre digit okt aaeeonccerem oeanice oe none right. The sample is, of course, too small for orobability statistics, but such a close coincidence on the part of one S seems rather striking. The tracer word “faint” did not appear in any of the statements, nor did Ross’ own tracer words. However, the S in Switzerland wrote in automatic writing a word which she inter- preted as “Amen.” Actually, the word written was ‘‘Ane.” Accord- ne to Mrs. a + ee yaaie he used to deliberately misspell r name, as ‘‘Ane.”’ This appears to be a ra i coincidence hovevee p ther improbable wo icated that Ross had been connected with the Rosi- grucians: the S in England stated this explicitly and the S in witzerland drew the Rosicrucian symbol of God's all-seeing eye. Ross had actually belonged to that organization, In contrast to these less significant correspondences, the most Linkage Experiments with Mediums 105 striking thing in the experiment was the number of apparent allu- sions to the locket incident. The locket was suspended on a fine chain and worn as a necklace. The S in Switzerland automatically drew chains in three variations and also something that looked like lace, such as a lace collar on the neck of a woman's dress, that is to say, “‘neck-lace” in a literal way. Her automatic writing also stated, “A graph, eine Tabelle, belongs to this drawing and will be reproduced in another place.” In other words, she was saying that the drawings were incomplete and would be supple- mented by another S, the basic idea of cross correspondence. There were no graphs in the scripts of the other Ss in this first iment, however. British subject also made a possible allusion to the neck- Jace, emphasizing lace: ‘‘Old-fashioned, with lace, bedspread, or something to do with lace. . . . It is a symbol, I think.” Then later, “I got jewelry and bits of gold things and material things in my Mrs. A commented on this statement: “The lace bedspread the English medium got was, I am sure, a beautiful coverlet that Ross’ mother had knit. I also have a number of other things she made—pillowcases, tablecloths, all with lace.”’ Thus, lace could have a sentimental meaning of its own for Ross and Mrs. A. How- ever, in the Swiss S’s automatic production “lace” and “necklace” seemed to be closely linked. It makes the British S’s statement that lace is a symbol (i.e, it stands for something aoe} much more meaningful. The possibility of real cross correspon in these statements can be singled out from chance coincidence only by secing the basic associative links that are ted. Athough these findings are interesting, the amount of cross correspondence is not as great as we would need for substantial evidence. However, Mrs, A got a “‘message” from Ross at the time of the experiment: “The letter is on its way—good news, but will find many delays before delivery, say, May first.” I felt that this statement might gy be a hint that the promised cross corres- pondence might be forthcoming in a subsequent session. Second experiment: In preparing for this experiment (in which there were two sessions), I wrote to Mrs. A and told her that to corroborate the findings of the first experiment it would be neces- to try to get similar results in a second series. In the letter giving her approval Mrs. A stated that Ross had su a certain e as a tracer. I thought this was too indistinct to be an effective stimulus and turned down the suggestion. Mrs. A persisted, however. In a letter dated April 6, 1956, she said, “I still think the perfume is a good target. I have a bottle that Ross gave me called ‘Golliwog.’”’

106 Journal of the American Society for Psychical Research An additional impetus to use the perfume as a tracer came in the same mail. I received a letter dated April 7, 1956, from the assistant in Philadelphia who was also serving as an S in which she discussed the steps to be taken for Ross’ next sessions. She said that she had had to interrupt the writing of the letter to me because *“... a wave of wonderful perfume suddenly surrounded me.” She interpreted it as “spirit perfume” given by one of her communi- cators. She said this was the first time such a phenomenon had occurred in her apartment and that investigation had shown that all her perfume bottles were tightly closed. We took this S's olfactory hallucination coinciding with Mrs. A’s letter as a good omen and decided to use the perfume tracer. If the hallucination meant anything in connection with the experi- ment, perhaps it was a hint that the S who had had the hallucination would be important in the next session. The other tracers used were the number 574 and the word “‘fallow-deer.” The next two experimental sessions took place on April 9 and 18, 1956. A total of eight Ss partici : one in Iceland, one in England, three in Philadelphia, and one in Texas; in addition, the two assistants in Philadelphia and New York also served as Ss (see Figure 5). Unknown to us at the time, Mrs. A had brought the Texas S into the experiment. When we heard of this, it seemed a senseless whim because this S had known Ross and had many memories of him which could have masqueraded as “messages.” However, this was not the case and, as will be seen below, she made a significant contribution. The instructions given to the Ss were the same as those for the first experiment (see pp. 103-104). Fic, 5. Links and participants in the second “Ross” experiment (April, 1956). Linkage Experiments with Mediums 107 Results: The results were again inconclusive. None of the Ss any of the experimental tracers. However, the New York Ss Gescribed Ross’ grave as being on the “northwest side of a hill, which was correct, according to Mrs. A. The New York S also gave Ross’ daughter’s name correctly and made strong allusions to Be scn-in-law'e occupation (involving photography and drawing). ‘The Texas S received the impressions “lace, handkerchief, cornish or coffin.” The lace idea appears close to the coffin asso- ciation in her automatic writing, as if they had been associated at The S in Philadelphia who had hallucinated the perfume usually produced “philosophical” messages by means of a pendulum over a circular arrangement of letters. To her great rise, she re- ceived (on April 9, 1956) the following com using that technique: Take r pencil and draw a line. Draw another line across the end. Halt of this line has been intersected or sree middle. Take another line, heavier, over the first line. Take another line across the end of the two — Now ae Veut cde Pisce ates side put a mark like a hinge, Repeat on the ri . Place a risi sun ea the side, and a hole in the middle of the sun. Place a long side knob at each end. Slow but sure. can be seen in Figure 6, the drawing the S made does not aut represent the commands. She told me later that she thought ¢ drawing was turning out to be a coffin and became very frightened of its symbolic significance, implying her own coffin, In the drawing she seems to have shaped the “long side knobs like coffin handles. Thus, there might be two aspects to be inter- reted: (1) A command to draw a box with two hinges, a keyhole ing a rising sun ornament, and two long side knobs, and (2) her own interpretation of the coffin and her emotions connected with it. Fro. 6. ‘Box-coffin” drawn by S in Philadelphia, When this quite definite piece of information emerged, I turned to Mrs. A for help in understanding the si ce of the drawing. In a letter dated Oesabier; 1962, she gave the following explanation:

108 Journal of the American Society for Psychical Research Yes, Ross did have a box resembling the one the Philadelphia subject drew—his mother also had one—but Ross had the necklace [locket] in his billfold when he was shot—I am sure he probably used both boxes at times .... I’m not an artist, but Ross’ box was oblong, leather, hand-tooled—I believe [made] in Italy. It had small compartments in it for rings, studs, cuff i and watch. He carried it with him in traveling and it was on his chest of drawers in To- ronto...” Apparently Mrs. A recognized the drawing as representing the box where Ross might have kept the locket. The drawing did not suggest a coffin to her, as it had to the Philadelphia S. Discussion of the Cross Correspondence Experiments If we hypothesize that the two sets of experiments show real cross correspondences in which the Ross personality tried to prove its survival after death, we must try to see which allusions repeat, reinforce, or clarify one another when compared. The Ss seem to have picked up references to the two spontaneous experiences which Mrs. A reported having had prior to the time of the experiments; the locket incident and the dream in which a box of orchids turned into Ross’ coffin. Fic, 7. Neckince-chain-lace cross correspondences. The necklace (locket) cross correspondences (shown in Figure 7) seem to be the most numerous, occurring prominently in the first experiment. Some references to this incident were oblique rather than direct; most allusions were to “chain,” “jewelry,” or “lace,” which has a syllable in common with “necklace.” Linkage Experiments with Mediums 109 In the first experiment the Swiss S made sketches resembling three chains, lace, and a lace collar. These seemingly i elements become significant when related to Mrs. A’s with the necklace, keeping in mind the prelogical nature of psi imagery. Chains are an obvious common element between the two events: both lace collars and necklaces are worn around the neck and, as cer above, oe and “necklace” share a syllable in common. English S’s statement emerging in frst experiment that “‘lace’’ is a symbol for something else is not immediately clear, since lace is not a commonly accepted symbol for anything else. However, if “symbol” is to mean only ‘a term with a shared significance for Ross and Mrs. A, then it is clear that the common le in “lace” and “necklace” could again be the association which could make “lace” a symbol for ‘‘necklace.” This connection is bolstered by the same elements appearing in the Swiss S's message. The English S’s additional impressions of owed and gold seem to reinforce the connection with necklace n her message. Mrs, A’s edahitton dream of the box of orchids which turned into Ross’ coffin also seems to have stimulated a series of cross correspondences (see Figure 8). In this series of allusions, references were more ightforward. For example, one S got “‘coffin’’ and another made a drawing which she identified as a coffin. Mar, "Sh Mon. A devume thos comde teen of crchide which ture ixte his culfle. Dinter! te KO. 1962. In the first experiment the Swiss S had stated that a graph “will be reproduced in another place.” This might also have been an indication of the relevance of the box ronnie which emerged during the second experiment; the possibility of a cross correspondence between these two messages is perhaps not far- fetched because the drawing is schematic enough to qualify as a graph.

110 Journal of the American Society for Psychical Research Although the second experiment seemed unproductive at first glance because none of the Ss mentioned any of the experimental tracers, the coffin-box drawing and the ‘‘coffin” and “grave” ref- erences which emerged are noteworthy. However, they do not seem to be closely related to Ross’ individual personality or his life events. The Texas S received the impression ‘‘coffin” preceded by “lace.” In her message alone the two parallel series of associations are joined. For this reason her contribution is significant. It was the Philadelphia S, however, who produced the most concrete picce of information during the second experiment: the box drawing. As we have noted previously, Mrs. A said that the drawing resembled Ross’ jewelry box where he might have kept the necklace. The box drawing, therefore, becomes related to the necklace motif. The Philadelphia S$ thought the box drawing represented a coffin and, as already noted, the Texas S's message made a connection between necklace and coffin. In this way, both motifs, necklace and coffin, are associated with the drawing of a box. These associations seem to parallel Mrs. A’s dream of March, 1953, in which a box containing something precious (orchids) turned into Ross’ coffin. In addition, during the second experiment the New York S got a message indicating that Ross’ grave was on the northwest side of a hill, a fact confirmed by Mrs. A. A grave is not a coffin, but could be an allusion to it, To get a complete picture of the complex nature of the cross correspondences, we must consider graphically not only the sepa- rate necklace and coffin motifs, as shown in Figures 7 and 8, but also the two motifs together (Figure 9). Besides the direct references which connect the two experiments and form the cross correspondences, there were three events which occurred shortly before the start of the second experiment and which can perhaps be interpreted as an attempt on the part of the Ross personality to point to the forthcoming cross corres- pondences: (1) Mrs. A’s communication from Ross about the delayed letter with good news seemed to me to indicate that another contribution to the necklace motif was to come later. The box drawing which emerged in the second experiment could possibly have been the fulfillment of that promise. The arrival in the same mail of (2) Mrs. A's letter suggesting (supposedly on behalf of Ross) that the perfume be used as a tracer and (3) the Philadelphia S’s letter 5 the hallucination of perfume provide a striking coin- ci - Because of it our attention was directed to the messages produced by the Philadelphia S, although her previous ones had Linkage Experiments with Mediums 111 been uninteresting. This proved to be a fruitful lead because she was the S who produced the most valuable material in the second experiment. Ape, "yl Mew. A tdieesifies drewing ae prowiry tee ie which Rens dept mechiacr, Fic. 9. Interlocking of necklace and coffin motifs. Various other hits and near-hits made by the Ss must be taken as a background of success against which to evaluate and inter- pret the cross correspondences. There were, for example, the isolated references to the Rosicrucians in the messages of the Swiss and English Ss in the first experiment; the Swiss S drew the Rosi- crucian symbol and the English S stated that Ross was a member of the group. Although these messages seem unrelated to other lines of association given by the Ss, they are related to facts about Ross’ life. Moreover, one S reproduced two of the three digits in the first tracer number. Another S wrote “Ane,” said to be Ross’ special mi ing of Mrs. A's given name. A third S gave the name of Ross’ daughter and came close to getting his son-in-law’s occupation. Tentative Interpretations The same difficulties involved in interpreting the classic cross correspondences of the British automatists in the S.P.R. group

112 Journal of the American Society for Psychical Research remain in the Ross case also. The two main alternatives to the survival hypothesis are the following: 1. The incidents are only a series of strange coincidences into which I projected meanings in much the same way as subjects do with Rorschach Ink Blots. 2. The incidents are, in spite of the linkage blocks, examples of ESP between living persons connected either with Mrs, A or with me; that such ESP accounts for the chain of associations found. Since the Ross cross correspondence experiments were not geared to quantitative analysis, the chance hypothesis cannot be evaluated. My own estimate, however, is that the coincidences are too numerous to be easily attributable to chance. The linkage design of the experiments decreases the plausibility of the hypothesis of telepathy among the living, especially tele- pathy from Mrs. A, which would be essential in explaining the coffin-box association in terms of ESP among the living. The participants were separated by the linkage design as shown in Figures 4 and 5. Mrs. A was geographically very remote from the main participants in the linkage chain, and except for the Texas S she did not know either the assistants or the Ss and they did not know her. The Swiss and English Ss who contributed the lace- necklace associations were three links removed from Mrs. A. I knew Mrs. A and the assistants in Philadelphia, New York, and Munich; I did not know of the participation of the Swiss, English, or Texas Ss until after the completion of the experiments. I also did not know the sub-assistant in London who formed one more additional link. Did the links we used really eliminate information about the DAs (the biologist and Ross) in the Ss’ statements? Important information appeared in the statements of the “far"’ Ss (Swiss, English) who were separated from the LAs by several links as well as in statements of the “‘near"’ Ss (Philadelphia, New York) whom I knew. It looks as though the introduction of more links did not eliminate the occurrence of cross co: dences and veridical statements about the DAs; whether the links decreased such cross correspondences and statements is not ascertainable without quantitative evaluations. Did the introduction of several links really exclude telepathy between LAs and Ss? The information is too scant for a definite answer to this question, However, no information concerning the LAs emerged in the linkage sessions which could not have been drawn from the DAs’ memories. It is true that some questions were asked about the LA (biologist’s widow) in the first pilot experiment, ¢.g., date of LA's birthday, and some approximately correct answers were given. However, the material was well within * Professor Linkage Experiments with Mediums 113 the range of the biologist’s memories, ¢.g., his birthday was on the same day as his wife's. In the Ross experiments re emerged in the Ss’ statements connected with the life of the L A) except items which were vitally concerned with Ross return of necklace, coffin dream). However, our instructions sen- sitized the Ss only to Ross and said nothing about the LA. To sum up, when the experiments were completed in 1956, my conclusion was that the results were not strong yon ® to conviction that communication with the deceased dsuired, _shejdlocery can dehiede eienibiciten tes ctr ter He bases additional and improved efforts in experimentation with the linkage design. Thus I began a massive exploration to help in building a 59 linkage experiment by sampling nearly a hundred mediums talent, as well as testing sitters and psychometric objects. In these explorations various assessment methods were devised and tried out such as paired-comparison designs, content analyses, position effects. These efforts paved the way to the next experiment. Tue Main LinkaGe Experiment? After conducting the exploratory tests described above involving linkage blocks, the appointment method, and cross co nd- ences, we used the knowledge we had gained to construct a formal experiment on survival. We that by drawing on the past literature and on our own prelimi ience, we could arrive at a modern experimental design which would yield specific and reliable information on this problem. Experimental Design The main linkage experiment aimed at exploring two variables and their interactions: 1. The status (living or dead) of the agent being received by the 2. The effects of setting up a linkage block of several persons between the S and the LA. We reasoned that if all mediumistic communication is based on information which the S obtains by ESP from an LA and his surroundings, one would expect: experiment, ensen, Marian Nester, Richard Sheargold, ‘an Steen- burgh were assi i jous capacities. The efficient help of these persons e

114 Journal of the American Society for Psychical Research eT nets y Sirol ba: soceecetmtgntel fn: yottings coment i tion about the LA himself than in getting information from the LA’s memories of the DA. (b) That the linkage block which seems to decrease the flow of information about the LA would also decrease the flow of infor- mation about the DA. On the other hand, if the S obtains information by ESP from the surviving personality of the DA, then one would expect: (a) That the S would be at least as successful in receiving messages about the DA as he would be in receiving messages about the LA. (b) That use of a linkage block would decrease the flow of in- formation about the LA obtained by ESP between the LA and the S, but would not affect the flow of information about the DA because ESP between the S and the DA would not be cut down by the block. The experiment also incorporated a number of other variables such as the personalities and idiosyncrasies of all persons involved (experimenters, assistants, Ss, As), the psi capacity of the Ss, etc. While these variables were only indirectly important for our theoretical purposes, any one of them could have played a large part in determining the success or failure of the experiment. Because the experiment involved so many variables, we had to analyze the factors statistically to gain an understanding of the effects of individual variables and the interactions between them, Professor Turner devised a complex statistical analysis of variance for our use. Therefore, we had to have the Ss’ qualitative state- ments assessed in a manner which would give scores which we could use in this analysis of variance (see p. 118 below, and Appen- dix), After extensive consultation with Professor Turner, we concluded that scores obtained from a scaled paired-comparison technique would be most advantageous for our statistical analyses. Such a technique involves pairing each one of the readings with every other reading; this meant that each LA was asked to judge twenty-eight pairs at a time. He was required to choose which reading he preferred in cach pair, rating each one on a7-point scale on which minus 3 represented a strong preference for reading A, 0 represented no preference, and plus 3 represented a strong preference for reading B. The paired-comparison technique introduced limits to the number of LAs we could use without making the checking task too cumbersome for them. We decided that we could not have more than four LAs and ep their task within bounds, while still retaining a balanced block design (see below). Linkage Experiments with Mediums 115 Another limitation was the number of Ss available. Because we wished to control variation in the test situation to a degree a having a New York City assistant, Mrs. Marian Nester, conduct the sittings, the Ss had to live in the metropolitan area. More- over, out of the many Ss we pre-tested only four gave sufficient —_ of producing psi in an experimental situation. These itations were decisive in determining the experimental design. As stated above, we used a balanced block design in which each S and each agent (LA and DA) was tested under manipulation of the two main variables. Each S gave four readings for each agent, as follows: 1. For an LA with linkage barrier. 2. For an LA without linkage barrier. 3. For a DA with linkage barrier. 4. For a DA without linkage barrier. The only condition we could not balance out was the chrono- ag sequence of readings with and without linkage barrier. e “with linkage” section of the experiment had to precede the “without linkage’’ section in all sessions because once an S had access to information the direct way (without linkage) he could not then be asked to secure the same information the indirect way (with linkage). Procedures To make the linkage block effective, each participant in the experiment had to be kept in ignorance about the other participants. We took precautions to insure this when we made our arrange- ments. As already stated, on the basis of previous testing we selected four New York mediums—A, B, C, and D—as Ss. Mrs. Nester, an assistant, was assigned the task of arranging and conducting the experimental sittings with these Ss. She was kept ignorant of the identity of the other assistants, Mr. 5 of London and Mrs, J of Dallas, Texas. These out-of-town assistants arranged sittings between LAs and “appointment mediums” (4 Ms) at their ends and supervised their LAs in checking the data in the later phase of the experiment. These assistants were asked not to tell their LAs who the experimenter (myself) was and they were not fo Mrs, Nester’s name. Each assistant arranged for two LAs to participate; their names were not given to me. The London LAs were close friends of Mr. S and the Texas LAs were close friends of Mrs. J. The linkage design of the experiment is shown in Figure 10.

116 Journal of the American Society for Psychical Research Fria. 10. Linkage design of the main experiment, In preparation for each test session with the Ss in New York, a preliminary “appointment” session was held with an AM in London or Dallas, At the appointment session the AM made an appointment with the DA. The AM, whether managed by Mr. S or Mrs. J (who was herself an AM), gave a private reading to the LA, then asked the DA if he would like to communicate through a very good S who would be concentrating on his name at a speci- fied future day and hour. As a rule, two DAs were lined up for each test session; twenty minutes were allowed for each. The LAs were instructed to do nothing in connection with the experiment on the day of the session, e.g., not to think about the DAs. At the predetermined times, Mrs. Nester held the sessions with the Ss in New York. All statements were tape-recorded. Blocked sittings: At these sittings Mrs. Nester gave the Ss only the names of the distant LAs and the DAs, alternating the order in different sittings. First and middle names or nicknames were used; last names were never given. The Ss then gave their im- pressions concerning the LAs and the DAs. Unblocked sittings: Sittings without the linkage design were based on the use of psychometric objects. Although researchers do not understand how or why such phenomena take place, they do know that if an S holds an object belonging to a distant A, the S's ability to reach that A by ESP is improved. In our experiment the Linkage Experiments with Mediums 117 ometric objects were used at both ends as focusing points the ESP connecting the Ss and the As. We assumed that this might help to offset the difficulties in finding the A which could be caused by the elaborate linkage arrangements. The procedure for the unblocked sittings was as follows: Mrs. Nester had made psychometric objects of her own available to the AMs in London and Dallas through the assistants in these cities. It was hoped that the physical presence of the psychometric objects would facilitate ESP contact hanbena Mrs. Nester and the AM. Each LA had provided Mrs. Nester with a psychometric object of her own and another belonging to the DA. Mrs. Nester gave the S these two psychometric objects at the beginning of the sitting. The objects were wrapped to conceal their nature. Paired-comparison judging: After each session the Ss’ statements were submitted to the LAs for paired-comparison judging. The LAs evaluated a set of twenty-eight pairs at a time without knowing which readings were intended for them. In this way eight sets of scaled evaluations were obtained from each LA, totalling thirty- two sets in all (see pp. 114 above, and Appendix). These data were submitted to the statistical analysis devised by Professor Turner, Results Statistical analysis: As shown in Table 1, most of the mean squares produced in the statistical analysis are on a level with error variance, thus indicating no significant phenomena. Psi effects are indicated only in the interactions with the item “Experi- mental versus Control” statements, which represent the difference between LAs’ ratings of statements intended for themselves as compared to those intended for other As. As can be seen in the table, when this variable was analyzed by itself, the results showed that in general the statements made by the Ss concerning a par- ticular LA did not seem to that LA to be any more appropriate to him than the statements which were not intended for him. When this “Experimental versus Control” variable was analyzed in interaction with other factors, only one three-variable interaction showed up as marginally significant: the interaction involving the living status of the As. Although this interaction may be simply due to chance, the hypo- thesis that it represents a real psi effect seems to be strengthened by the high mean squares produced by two other interactions deal- ing with the As’ living status. Interaction between the living status variable and the presence or absence of linkage blocks gave mean

118 Journal of the American Society for Psychical Research squares about three and a half times the error variance. Interaction among the three variables of living status, blocked or unblocked treatment, and identity of the As also gave mean squares which are high enough to be statistically significant. This indicates that idiosyncrasies of the As interfered either in their rating of the paired comparisons or with the Ss’ success in making statements about them. Table | ANALYsis OF VARIANCE OF ParneD-ComPanison Scores Source of Variance “Whole Plots"*t Between Between t Status (LA va. DA) Error Between Agents (LA & DA)* Between Treatments (Blocking va. Agents (LA & DA) x Treatments Subjects x Agents (LA & DA)* Steet : Agents (LA & DA) x x Agent Sta ‘LA DA) 2 t Status Agents (LA & DA) 3 t Status (LA vs. DA) x rs Agent Status (LA vs, DA) x Agents (LA & DA OTNeimens 3.40 10 3.24 .025 an Linkage Experiments with Mediums 119 Inspection of data: Besides looking at the mean squares produced by the analysis of variance, we also inspected the data. Al this cannot yield statistically valid evidence, we felt that it might ive us clues as to what went on during the experimental sessions. € must emphasize that none of the trends we found were ex- pressed strongly and uniformly enough to appear in the variance table and they may all have been due to chance. Inspection of the data revealed that on the whole the Texas LAs _ were much more successful than those in London both in identify- ie which were related to themselves and in identifying those ted to their DAs. Inspection also revealed that one of the Ss, C, had much more success than the others, She was quite successful in all four con- ditions; blocked and unblocked, with living and deceased As. Subject A had success only in blocked experiments with LAs and slight success with un ed experiments and DAs. Subject B ormed best in unblocked experiments and with DAs, having only slight success with LAs in blocked sessions. Subject D had good success only with unblocked sessions and DAs. In all other situa- tions she seemed to exhibit strong psi-missing tendencies; her scores were below chance expectation. Inspection of the data also indicated psi-missing effects in some cells of the data of subjects A and B, Discussion When designing the experiment we anticipated uniform trends with few strong idiosyncrasies among the As and Ss. We also expected a considerable amount of straightforward psi hitting, with scores varying from mean chance expectation to a positive deviation. In reality, there seemed to be a large variation in relation to the personality characteristics of the Ss and of the As, and some evidence of psi missing. These factors appear to have prevented uniform trends in the data. The differing scoring patterns of the Texas and English LAs mentioned above—the Texans giving higher scores when iden- tifying items concerning their DAs, the English LAs scoring posi- tively only when identifying items concerning themselves—may have been due to ESP or to chance, Assuming that an ESP effect was involved, we could interpret the variance obtained to mean that some LAs have cither more ability or more motivation to correctly identify items related to themselves, while others have more ability or motivation in the case of items relating to their DAs. This trend as it occurred in our data could also be explained on the assumption that the Ss reacted differently to the Texas and English LAs and DAs. We have no basis for discriminating between these possibilities.

120 Yournal of the American Society for Psychical Research The indications of psi missing in the results of three out of the four Ss were surprising. The analysis of variance design used was not well equipped to deal with psi-missing phenomena inter- mingled in haphazard fashion with psi hitting. This might have been one reason for the low level of results obtained. The concept of psi missing—the theory that the S uses ESP to “turn away” correct responses—has not been explored in any studies of medium- ship of which we are aware. On the basis of our experience we believe that it should be considered in experimental design as well as in theory. It would also seem that the statistical analysis design selected was not well adapted to detecting the sources of variance as they were actually encountered. For this reason, research designs should incorporate factors not only capable of detecting psi missing, but also idiosyncratic differences among participants. Or, to take a different route, constructing experiments in which a larger num- ber of As and Ss are involved might be another way of overcoming the difficulties. In general, it seemed to my colleagues and myself that the flow of psi imagery became more inhibited as the experimental situation was made more structured and artificial. The relatively unsophisti- cated preliminary experiments seemed to yield more definite ESP effects than the main experiment based on modern methods. Although the main experiment did not yield much evidence of straightforward psi hitting, it seemed worth putting on record in detail because it throws light on possibilities for future research in mediumship: on what to do and what to guard against doing. The computer age does offer new techniques for survival research. Extremely powerful methods are now available to separate factors hitherto deemed inseparable. We used one method—multi- variate experimental design—in a bold venture into a little-known territory. In spite of extensive explorations, we did not manage to fit the new tools well enough to the old phenomena; for example, we did not anticipate marked psi-missing effects in mediumistic communications. However, this sort of “failure” is often the price one has to pay for radically new approaches in research, We do not believe that any of the difficulties are insurmountable, On the contrary, the rich yield of our exploratory experiments indicates that classical phenomena such as cross correspondences are still obtainable in our times. New machines certainly can be adj to grind the type of grain fed into them. They are capable of separating such entangled, but vital, factors as ESP between the living from messages ostensibly originating from the deceased. This can be done, but in my estimation long-term research efforts on the part of many investigators will be necessary. Linkage Experiments with Mediums 121 REFERENCES ]. Auuison, Lypta W. “Proxy Sittings with Mrs. Leonard.” Proc. S.P.R., Vol. 42, 1934, 104-145. 2. Batrour, G. W. “The Ear of Dionysius: Further Scripts Affording Evidence of Personal Survival.” Proc. S.P.R., Vol. 29, 1918, 197-243; 260-286. . “A Study of the Psychological Aspects of Mrs. Willett’s Mediumship, and of the Statements of the Communi- cators concerning Process.” Proc. S.P.R., Vol. 43, 1935, 41-318. . Barrett, W. F. Death-bed Visions. London: Methuen, 1926. . Broap, C. D. “The Phenomenology of Mrs. Leonard’s Mediumship.” Journal A.S.P.R., Vol. 49, April, 1955, 47-63. 6. CaRINGTON, Waarecy. “The Quantitative Study of Trance Personalities.” Proc. S.P.R., Vol. 42, 1934, 173-240; Vol. 43, 1935, 319-361; Vol. 44, 1936-37, 189-222; Vol. 45, 1938-39, 223-251. 7. Dovps, E. R. “Why I Do Not Believe in Survival.” Proc. S.P.R., Vol. 42, 1934, 147-172. 8. Ducasse, C. J. A Critical Examination of the Belief in a Life After Death. Springfield, Minois: Charles C Thomas, 1961. 9. EHRENWALD, Jan. New Dimensions of Deep Analysis: A Study of Telepathy in Interpersonal Relationships. London: Allen & Unwin, 1954. 10. Gurney, E., Myers, F. W. H., and Popmorg, F. Phantasms of the Living. London: Tribner and Co., 1886. 11. Hatt, R., and Trousripce, U. “On a Series of Sittings with Mrs. Osborne Leonard.” Proc. S.P.R., Vol. 30, 1918-19, 339-554. 12. Hart, Hornet. “Six Theories About Apparitions: A Coopera- tive Report with Associated Collaborators in the International Project for Research on ESP Projection,” Proc. S.P.R., Vol, 50, 1956, 153-239. . The Enigma of Survival: The Case For and Against an After Life. Springfield, Illinois: Charles C Thomas, 1959. 14. Hopcson, Ricuarp. “A Record of Certain Phenomena of Trance.” Proc, S.P.R., Vol. 8, 1892, 1-167. 15. James, WitttaM. “A Record of Observations of Certain Phe- nomena of Trance.” Proc. S.P.R., Vol. 6, 1889-90, 651-659. 16. Lopce, Ottver. “Evidence of Classical Scholarship and of Cross-Correspondence in some New Automatic Writings.” Proc. S.P.R., Vol. 25, 1911, 113-175. 17. Mureny, Garpner. “An Outline of Survival Evidence.” Journal A.S.P.R., Vol. 39, January, 1945, 2-34. us

122 Fournal of the American Society for Psychical Research 18. ———. “Difficulties ing the Survival Hypothesis.” A.S.P.R., Vol. 39, April, 1945, 67-94. 19. ———. “Field Theory and Survival.” Journal A.S.P.R., Vol. 39, October, 1945, 181-209. 20. —__——. ‘Triumphs and Defeats in the Study of Mediumship.” Journal A.S.P.R., Vol. 51, October, 1957, 125-135. 21. - Challenge of Psychical Research: A Primer of Para- psychology. New York: Harper & Brothers, 1961. 22. Myers, F. W. H. Human Personality and its Survival of Bodily Death. London: Longmans, Green and Co., 1903. 23. Osis, Karuis, ‘‘Deathbed Observations by Physicians and Nurses.” Parapsychological Monograph No. 3. New York: Parapsycho Foundation, Inc., 1961. 24. Pippincton, J. G. “A Series of Concordant Automatisms,”’ Proc. S.P.R., Vol. 22, 1908, 19-416. . Pratt, J. G. “Towards a Method of Evaluating Mediumistic Material.” Bulletin23, Boston Society forPsychic Research, 1936. 26. Pratt, J. G., and Birce, W. R. “Appraising Verbal Test Material in Parapsychology.” Journal of Parapsychology, Vol. 12, December, 1948, 236-256. 27. Rune, L. E. “Hallucinatory Psi Experiences: I. An Intro- ductory Survey; II. The Initiative of the Percipient in Hallu- 3 cinations of the Living, the Dying, and the Dead; III. The Intention of the Agent and Dramatizing Tendency of the Percipient.” of career okey, Vol. 20, December, 1956, 233-256; Vol. 21, March, 1957, 13-46; Vol. 21, Sep- tember, 1957, 186-226. . “Spontaneous Physical Effects and the Psi Process.” Journal of Parapsychology, Vol. 27, June, 1963, 84-122. 29. Rout, W. G. “The Contribution of Studies of ‘Mediumship’ to Research on Survival After Death.” Journal of Parapsychol- ogy, Vol. 24, December, 1960, 258-278. 30. Savrer, W. H, Ghosts and Apparitions. London: G. Bell & Sons, Ltd., 1938. 31. Satrmarsu, H. F. Evidence of Personal Survival from Cross Correspondences. London: G. Bell & Sons, Ltd., 1938. 32. Sattmarsn, H. F., and Soar, S. G. “A Method of Estimating the Supernormal Content of Mediumistic Communications.” Proc. 5.P.R., Vol. 39, 1930-31, 266-271. 33. Sipewick, E. M. “A Contribution to the Study of the Psychol- ogy of Mrs. Piper's Trance Phenomena.” Proc. S.P.R., Vol. 28, 1915, 1-652. - “On the Development of Different Types of Evidence for Survival in the Work of the Society.” Proc. 8.P.R., Vol. 29, 1917, 245-259. Linkage Experiments with Mediums 123 35. ———. “Phantasms of the Living ..."” Proc. S.P.R., Vol. 33, 1923, 23-429. 36. Stevenson, IAN. “Twenty Cases Suggestive of Reincarnation.” Proc. A.S.P.R., Vol. 26, 1966. (In press.) 37. Tuomas, C. Drayton, “A Proxy Ease Extending over Eleven Sittings with Mrs. Osborne Leonard.” Proc. S.P.R., Vol. 43, 1935, 439-519. 38. ———. “A Proxy Experiment of Significant Success.” Proc. S.P.R., Vol. 45, 1939, 257-306. 39. Tuomas, JouN F. Beyond Normal Cognition: An Evaluative and Methodological Study of the Mental Content of Certain Trance Phenomena. Boston: Bruce Humphries, Inc., 1937. 40. Tyrrett, G. N. M. Apparitions. London: Gerald Duckworth & Co., 1943. 41. Waker, Nea. Through a Stranger’s Hands: New Evidence for Survival. London: Hutchinson & Co., 1935. A.S.P.R. 880 Fifth Avenue New York, N. Y. 10021 APPENDIX Judging Procedure All repetitions and items of an unverifiable nature were edited out by a person having no knowledge of the correctness of the statements before the readings were submitted to the agents for judging. : Each agent was sent eight peckagse of paired readings to judge, one package at a time. Each package contained eight readings from one medium, The four “with linkage” (blocked) readings were randomly mixed with the four “without linkage” (unblocked) readings. Two of the readings (one blocked and one unblocked) concerned the particular agent herself. The remaining six readings were intended br the other three agents and so constituted controls. The readings were arranged in twenty-eight combinations of two readings, each pair in a numbered envelope. The pairs were compared regardless of their order, i.e,, whether A or B came first. If, for example, A had to be compared with B, the question would not later be asked about B in comparison with A; instead, the score of this comparison would be computed by reversing the sign for rating A versus B.

124 Journal of the American Society for Psychical Research This feature of the paired-comparison technique made each judgment scaling matrix of fifty-six judgments symmetrical, i.e., the total score of each matrix is zero. Each score is used twice, with opposite signs. This feature did not interfere with the purpose of the experiment because the main interest was in comparing the differences between the two classes of statements: (1) “experi- mental statements” (those meant for the particular sitter or his communicator) and (2) ‘‘control statements’? (those meant for other sitters or their communicators). The symmetrical nature of the paired-comparison matrix renders some items in the analysis of variance table zero. PRINTED IN BRUGES, BELGIUM, BY THE ST CATHERINE PRESS, LTD.