Leonard M. Wagman, Atty., Marcel Mallet-Prevost, Asst. Gen. Counsel, Dominick L. Manoli, Associate Gen. Counsel, Arnold Ordman, Gen. Counsel, Warren M. Davison, Atty., N.L.R.B., Washington, D.C., for petitioner.

L. G. Clinton, Jr., H. L. Deakins, Jr., Houston, Tex., Fulbright, Crooker, Freeman, Bates & Jaworski, Houston, Tex., of counsel, for respondent.

Before WISDOM and GEWIN, Circuit Judges, and HANNAY, District Judge.

WISDOM, Circuit Judge:

In this unfair labor practice case Camco, Inc., a Texas corporation manufacturing oil field equipment, contends that it made a good faith reduction in force because of a decline in business brought about by Hurricane Carla. The charging union, District Lodge No. 37, International Association of Machinists, AFL-CIO, contends that the decline in business was a pretext for discharging eleven employees because of their union activities. Camco insists that the eleven men were selected for termination of employment before it knew of their union activities. The Trial Examiner accepted the Company's explanation, found that Camco had engaged in certain 'technical' violations of the National Labor Relations Act but had not engaged in the other alleged unfair labor practices, and concluded that a cease and desist order would not 'effectuate the purpose of the Act.' The National Labor Relations Board disagreed with the Examiner's conclusions. The Board found that the Company violated Section 8(a)(1) of the Act by engaging in an 'extensive campaign of interrogation coupled with promises of benefits and implied threats' and that this 'coercive conduct warrants the issuance of a remedial order.'1 The Board found that Camco violated Section 8(a)(3) and (1) of the Act by discriminatorily discharging the eleven employees. The Board seeks enforcement of an order based on its findings. We grant enforcement of the order, except as to the reinstatement of two of the discharged employees.

I.

Coercion by interrogation is one of the 'subtler' forms of management's interference with labor's protected rights.2 As the differences here between the Examiner and the Board illustrate, the 'act of interrogation is not coercive or intimidating on its face, nor is it easy to demonstrate just how often, and under what circumstances, a threat of reprisal will be inferred by the employees.'3 Bourne v. N.L.R.B., 2 Cir.1964, 332 F.2d 47, 48 modifying 144 NLRB No. 75 (Sept. 26, 1963), is helpful in determining the limits of proper interrogation. In that case the court lists five factors to be considered in weighing the lawfulness of company interrogation of employees:

'(1) The background, i.e. is there a history of employer hostility and discrimination?

'(2) The nature of the information sought, e.g. did the interrogator appear to be seeking information on which to base taking action against individual employees?

'(3) The identity of the questioner, i.e. how high was he in the company hierarchy?

'(4) Place and method of interrogation, e.g. was employee called from work to the boss's office. Was there an atmosphere of 'unnatural formality'?

'(5) Truthfulness of the reply.'

This list is not intended to be definitive and, as Professor Bok has pointed out, intimidation may occur even if all of these factors cut in favor of the employer. He warns that 'employers must beware of interrogation unless (1) they have a valid purpose for obtaining information concerning the union's strength; (2) they communicate this purpose to the employees; and (3) they assure the employees that no reprisals will be taken.'4 Taking the Bourne approach, but bearing in mind Bok's formulation of the Board's warning to employers, we enter a Blus Flash5 thicket.6

There is no doubt about Camco's antiunion animus, although there is contradictory testimony in regard to a speech by the Company's president in which the anti-union policy was plainly stated. Several employees testified that in October 1961 President Harold E. McGowen, Jr. made a speech to the employees in which he threatened to close the plant if it were unionized. McGowen testified that he made no such threats. His speech was in fact a written statement of over-all plans read at a meeting of stockholders as well as at a meeting of employees. The Examiner discredited the testimony of the employees. The following anti-union statement does appear in the written speech:

'Union-- as you know, some 3 1/2 years ago, we had a union election in this plant. This was defeated and since then, we have had absolutely no trouble in this respect. This Company is of the very firm opinion that we will not have a union here. We are against the union in all respects and will do our best to see that it does not come to this plant. If any man feels otherwise, we invite him to leave and go to another plant. This is the American way of life.'

In December 1961, President McGowen decided to reduce production personnel because sales and profits were down. Camco terminated four employees on January 4, 1962, six on January 5, and eight more later in the month. February 2, Hughes, head of the production department, instructed Shop Superintendent Theek to submit a list of ten men to be laid off. Theek passed the instructions along to his foremen, Cook and O'Pry. February 7 or 8, the foremen submitted to Theek ten names. All were union supporters and all were terminated February 16.

Efforts to organize the plant did not begin until the middle of January when McCall, a production employee asked the help of the Union. February 9, Locke, another production employee, arranged an organization meeting at the Union Hall the next day.

Numbers are important in this case. Of the 95 employees in the production department, 16 attended the meeting and signed authorization cards. Of the 16, Camco fired 11; one on February 13, three days after the meeting, and 10 on February 16. These 11 men were the only production employees discharged during February 1962. No non-union man was discharged.

Between the time of the meeting and the time of the terminations, three Camco officials (Shop Superintendent Theek and Foremen O'Pry and Cook) interrogated 11 employees about their union activities; 10 were union adherents. The Company admittedly interrogated nine of the eleven union employees discharged; the tenth was a union leader discharged summarily; there is no evidence one way or the other as to the eleventh.

The discharged employees testified that the interrogators, Theek, O'Pry, and Cook attempted to learn the identity of the men who had been at the meeting and repeatedly said that the men who had attended the union meeting would be fired. Cook testified that the Company knew the identity of the men.

The testimony of the discharged employees was corroborated by the testimony of a union sympathizer, Baggett, who was still working for Camco at the time of the hearing. He testified that when he was questioned he did not admit his union activity; his foreman testified that he did. According to his testimony, one Camco foreman stated that he would not be fired because he had not attended the union meeting. Another foreman stated, before the union meeting, that he was going to the union hall to take names. A foreman said shortly after the discharge of the ten men that they had been fired because they were 'troublemakers and union agitators'.

Superintendent Theek admitted that he interrogated one employee. Foreman Cook admitted interrogating four employees, including the one Theek questioned. Foreman O'Pry admitted interrogating seven employees. These interrogations took place in a short period-- between February 13 and February 15. February 15, Theek asked employee Locke if he had attended the union meeting and mentioned the benefits that the Respondent had given the employees. February 13, Cook asked Locke whether he had heard rumors 'going around' about a union. He told Locke 'to keep his feet on the ground' and that it would be to Locke's benefit if he 'stayed out of it'. During this same period, O'Pry, on separate occasions, asked employees McCall, Clepper, Cox, Yeager, and Williams whether they had attended the union meeting.

According to the testimony of Camco supervisory employees, at least 6 of the 10 union adherents questioned denied that they knew anything about the union activity or that they had attended the union meeting. A seventh testified that he denied it; his foreman testified that he admitted it. In N.L.R.B. v. Syracuse Press, 2 Cir.1954, 209 F.2d 596, 597, cert. denied, 1954, 347 U.S. 966, 74 S.Ct. 777, 98 L.Ed. 1108, discussed in N.L.R.B. v. Harbison-Fischer Mfg. Co., 5 Cir.1962, 304 F.2d 738, 744, the Second Circuit emphasized the fact that one of the employees when questioned concealed his union activity. Here, since most of the Camco employees questioned, at least at first, concealed their union activities, the Board could reasonably infer that they were answering under pressure.

Camco argues that the interrogation was only by low-ranking officials and was done informally in the shop while the men were at work. Theek supervised the machine shop. Cook and O'Pry were foremen under Theek. The place of interrogation and the rank of the interrogators bear on the issue of coercion but, assuming the authority of the interrogators to speak for the Company, the crucial question is not their rank but whether to the employees the interrogators represent the Company. See Hendrix Mfg. Co. v. N.L.R.B., 5 Cir.1963, 321 F.2d 100. Here the record shows that in the minds of the production employees the foremen and supervisor did indeed represent Camco management. Because of the nature of the interrogation and the obvious correlation between the interrogation and the discharges, the Board could reasonably conclude that the questioning was not casual, whether it took place in the shop or in the office of the President of the Company and whether President McGowen or Superintendent Theek was the inquisitor.

Bourne points out that an innocuous question (e.g. How is the union doing?) is much less coercive than a pointed question (e.g. Who are the union leaders?). 332 F.2d at 48. Two of the Camco employees who answered untruthfully refused to answer innocuous questions (Did they know anything about the union rumors?). If an employee refuses to give a truthful answer even to an innocuous question, the inference of coercion is as strong as if he refused to answer a pointed question.

Questioning is much more likely to have a coercive effect if the purpose of the interrogation is not explained and if there are no assurances against retaliation. See Martin Sprocket & Gear Co. v. N.L.R.B., 5 Cir.1964, 329 F.2d 417. Camco made no explanation to the interrogated employees and gave no reassurances that their answers would not be held against them. Nothing was done to negate the effect of questioning most of the union adherents, or the effect of Camco's known anti-union policy.

The evidence of promised benefits is not conclusive. The Company made no promise of any specific benefit, but the testimony of Theek, Cook, and O'Pry shows that they did make an effort to persuade the men that it was against their interest to join the Union. Theek's conversation with Locke included a discussion of Camco's employee benefits as compared with the benefits from joining the Union. As for threats, Cook made certain statements to Bounds and Ledbetter which might properly be considered as 'implied threats'.

Camco deprecates any notion that the questions put to the employees were coercive. The Company draws attention to these facts. There were only eleven acts of interrogation, ten by minor supervisors. The other inquiry, by Theek, was made to an employee who had been his friend for thirty years. The interrogation consisted only in asking whether the employee had attended the union meeting or had heard about a union rumor, without any effort to obtain other information on union activities. Finally, so Camco contends, such remarks were natural and could be expected in an atmosphere in which such discussions were occurring 'all over the shop.'

The questioning cannot be plucked out of context. And the context includes these considerations. (1) The Company had publicly announced an anti-union policy. (2) The law of probabilities indicates a close connection between the selection of men for interrogation and the union activities of those men. (3) There is a manifestly close correlation between the interrogations and the terminations of employment. (4) The interrogation was systematic and intensive over a period of only three days. (5) Some of the interrogatees gave untruthful answers, apparently for fear of the consequences of revealing their union activities. (6) The Company did not explain the purpose of the interrogation to the employees and gave no assurances against reprisals. (7) There is some evidence of coercion by carrot and by stick.

We hold therefore that fair inferences and substantial evidence compel affirmance of the Board's finding that Camco's interrogation of its employees violated Section 8(a)(1) of the Act.

II.

The issue of discriminatory discharge has four distinct components: (1) the employer testimony, (2) the employee testimony, (3) the Examiner's findings, and (4) the Board's inferences from the admitted facts of the case. Here, Camco management testified that the discharges could not have been discriminatory, because Camco decided to discharge the men before the union activity even began. The employees testified that Camco's supervisor and foremen told them that men were going to be fired because of their union activity. The Trial Examiner discredited the employee testimony and credited the employer testimony. The Board, largely relying on inferences from admitted facts, held that the men whose employment was terminated were discriminatorily selected for discharge because of their union activities. Each of these four components creates its own problems.

In earlier decisions in this circuit, we have said that because reinstatement of a discharged employee may be considered a penalty, the evidence of discrimination should be compelling. We said, too, that the Board should accept the employer's testimony if the evidence gives equal support to the employer's explanation and the inference of discrimination, since the employer, testifying under oath, is in a position to know the motive. N.L.R.B. v. Houston Chronicle, 5 Cir.1954, 211 F.2d 848; N.L.R.B. v. Tex-O-Kan Flour Mills Co., 5 Cir.1941, 122 F.2d 433. However, many times, too often to cite, we have stated and restated the substantial evidence rule as phrased by the Supreme Court in Universal Camera and Walton. And, at least since our decision in N.L.R.B. v. Coats & Clark, 5 Cir.1956, 231 F.2d 567, 572, we have held that we cannot disturb the Board's choice if there is a fair conflict between the employer's testimony and a reasonable inference of discrimination. See N.L.R.B. v. Walton Mfg. Co., 5 322 F.2d 187, 189; N.L.R.B. v. Florida Citrus Canners' Cooperative, 5 Cir.1963, 311 F.2d 541. In N.L.R.B. v. Walton Mfg. Co., 1962, 369 U.S. 404, 409, 82 S.Ct. 853, 7 L.Ed.2d 829, the Supreme Court eliminated any residual doubt left over from Tex-O-Kan and reaffirmed the tenet that the employer's reasons for his conduct in a discriminatory discharge case are to be given no more weight than in any other type of case under the Act: 'There is no place in the statutory shceme for one test of the substantiality of evidence in reinstatement cases and another test in other cases.' 369 U.S. at 407, 82 S.Ct. at 855, 7 L.Ed. at 832.

Still, under Walton and Universal Camera Corp. v. N.L.R.B., 1950,340 U.S. 474, 71 S.Ct. 456, 95 L.Ed. 456, we must give the employer's testimony whatever weight it deserves in the light of the record as a whole. The employer testifies under oath, and the employer is the only one in a position to know why an employee is discharged. The employer's testimony cannot be lightly disregarded. N.L.R.B. v. Florida Citrus Canners' Cooperative. However, an 8(a)(3) violation always depends on the employer's intent. Those who testify to the circumstances by which the elusive fact of intent must be proved testify under the same oath used by the employer. The employer's oath carries no more weight than an employee's oath.

Universal Camera spells out the role of the Trial Examiner. There, the Board reversed the Trial Examiner; the Court of Appeals disregarded his findings. The Supreme Court held that the Board's findings must be affirmed if substantial evidence in the record as a whole supports the Board, but the Court also made it clear that the findings of the Trial Examiner are a part of the record, and must be given whatever weight they merit in the circumstances of each case.7

When the Board and the Trial Examiner disagree, the Board takes into account the evidence discredited by the Trial Examiner as well as other evidence; it is still part of the record. Similarly, this Court takes into account the evidence relied on by the Trial Examiner as well as the evidence relied on by the Board. The Board may rely on inferences in conflict with testimony the Examiner credits. And this Court may rely on the Board's inferences rather than testimony the Examiner credits. As the Second Circuit puts it:

'In reaching this conclusion he (the Examiner) relied largely on the testimony of * * * the Company's manager, as to what his intentions were. The Board, on the other hand, drew contrary inferences from the undisputed facts. Although the Board may not overrule its Trial Examiner by discarding the positive credible testimony of a witness in favor of an inference drawn from tenuous circumstances * * * it may refuse to follow its Trial Examiner in crediting testimony where it conflicts with well supported and obvious inferences from the rest of the record. Such refusal is particularly justified where the testimony in question is given by an interested witness and relates to his own motives.' N.L.R.B. v. Pyne Molding Corp., 2 Cir.1955, 226 F.2d 818.

One of the basic premises of Universal Camera is that, 'The findings of the examiner are to be considered along with the consistency and inherent probability of testimony'. 340 U.S. at 496, 71 S.Ct. 469, 95 L.Ed. 472. This principle is part of the trend towards more liberal rules governing evidence and methods of proof: 'However halting its progress, the trend in litigation is toward a rational inquiry into truth, in which the tribunal considers everything 'logically probative of some matter requiring to be proved." 340 U.S. at 497, 71 S.Ct. at 469. The trend includes inferences based on common sense recognition of probabilities.

Camco argues against the Board's playing a 'numbers game'. But courts have long accepted an inference of discrimination based on the percentage of discharged employees involved in union activities. Courts should. A man cannot cross a street without first having made a flash calculation of his chances of reaching the other side, based on his experience with the law of probabilities. In more legal language, in N.L.R.B. v. The Newton Company, 5 Cir.1956, 236 F.2d 438, 445-446, we said:

'Evidence is admissible in a proceeding to determine whether or not there has been anti-union discrimination in a lay-off of the percentage of union adherents involved. * * * The drawing of such inferences as are to be drawn from percentage testimony is within the province of the Board and if reasonably drawn, the inference is evidence, and with other evidence may be substantial evidence such as to entitle the Board's finding based thereon to our acceptance.'

In Newton the Court found that the inferences were overbalanced by other evidence in the case. But this Court and other courts have relied on such inferences as a basis for enforcing the Board's orders. See N.L.R.B. v. Nabors, 5 Cir.1962, 196 F.2d 272, 275-276 cert. den'd, 1952, 344 U.S. 865, 73 S.Ct. 106, 97 L.Ed. 671; N.L.R.B. v. Sandy Hill Iron & Brass Works, 2 Cir.1947, 165 F.2d 660, 663; N.L.R.B. v. Dinion Coil Co., 2 Cir.1952, 201 F.2d 484; cf. Stokeley Foods, Inc. v. N.L.R.B., 5 Cir.1952, 193 F.2d 736. If there is a choice between 'two fairly conflicting inferences', we must uphold the Board. N.L.R.B. v. Coats & Clark.

If we should consider only the testimony of the respondent and the employees and for the moment disregard the Examiner's findings and the Board's inferences, we should say that the evidence on both sides is about even. Camco's management and supervisors testified that they did not know the identity of the union adherents, and that they selected the men on February 8, before the union activity began. This testimony was internally consistent. It was not weakened to any extent on cross examination. According to the testimony of the discharged employees, the Camco foremen and the supervisor told them that Camco knew who the union adherents were, or would soon know. The Camco foremen and supervisor also said that the union adherents would soon be fired. Some but far from all of this testimony was weakened on cross examination. The discharged employees were corroborated by Baggett, a union adherent who was not fired and was still working for Camco at the time of the hearing. Only Baggett seems free of self interest.

If we add to the testimony the Board's inferences drawn from admitted facts, the record weighs very heavily against Camco. Camco bases its version of the facts on an incredible series of coincidences.

As noted in the discussion of the interrogation issue, ten of the eleven men questioned were union adherents. But there were only sixteen union adherents in the production department where the men worked. Thus, of seventy four nonunion production employees, only one was questioned. We have to say that it is a fair inference that Camco must have known the identity of most of the union adherents.

Of the sixteen union adherents, Camco discharged eleven. Camco discharged not one of the seventy four non-union employees. Again we have to say that it is a fair inference that Camco based the discharges on the discriminatees' actual union activity and Camco's knowledge of the fact the discriminatees favored organization of a union at the plant.

This only begins the chain of coincidences. There is a close correlation between the men questioned and the men discharged, whether we examine the totals or the number under each foreman.8

One further inference weakens Camco's explanation. Camco's management testified that some men were terminated in January to cut production costs in the face of lagging sales. The men were terminated a day or two after Camco made the decision. On February 7 Camco supposedly drew up a list of men to be discharged eight days later. Camco offered no evidence to explain the delay. The supposed eight days delay included the beginning and the end of the union activity at Camco. The Board could reasonably infer that Camco did not in fact select the employees before the union activity began.

Camco objects to the Board piling inference upon inference. The so-called rule against pyramiding inferences, if there really is such a 'rule' and if it is anything more than an empty pejorative, is simply legalese fustian to cover a clumsy exclusion of evidence having little or no probative value. See 1 Wigmore, Evidence (3d ed. 1940) 41. In labor discrimination cases the employer usually resorts to the rule when knowledge of union activity and discriminatory motive rest on the same bit of evidence. See e.g., N.L.R.B. v. Miami Coca-Cola Bottling Co., 5 Cir.1955, 222 F.2d 341; Interlake Iron Corp. v. N.L.R.B., 7 Cir.1942, 131 F.2d 129. Here, however, there is a broad base supporting the inference of knowledge and the inference of motivation. Call them 'findings' or 'conclusions' and they may be piled on inferences. Here, the inferences rest on many admitted facts: the total number of union adherents, the total number of production employees, the total number of union men and nonunion employees questioned, the total number of discharged employees questioned, the total number of union and nonunion employees questioned and discharged in each section of the production department, the unexplained lapse of time between the supposed decision to fire and the actual discharge of the men. There are other significant factual bases on which the Board rests its inferences. In short, the multiple inferences in this case have sufficient probative value to be weighed in the Board's and this Court's reaching a decision. See N.L.R.B. v. W. C. Nabors.

This case is well within the limits marked out by earlier cases. In N.L.R.B. v. Walton Mfg. Co. the Board's finding was supported by some employee testimony which directly conflicted with the employer's explanation, but the inference based on the percentage of men discharged was not as strong as the inference in this case, since only slightly more than half of the discharged employees were union adherents. In N.L.R.B. v. W. C. Nabors there were veiled threats to union adherents but there was no mass of testimony of direct statements that the union adherents would be fired. There was a strong statistical inference (twenty three of the twenty six men discharged were union adherents); in this case all of the men discharged were union adherents. In N.L.R.B. v. Dell, 5 Cir., 283 F.2d 733, as in the present case, all of the discharged employees were union adherents but a large percentage of the employees as a whole were not. We said:

'But it seems to us that it was more than a coincidence that each discharged employee was active in the Union organizing movement and each was discharged during the period when the Company was opposing the Union by threats, promises, and other unlawful acts of interference. The record fairly permits the conclusion that anti-union motivation lay behind the Company's action.' 283 F.2d at 736.

Most of the cases denying enforcement lacked at least one of the factors pointing to discrimination in this case, or there were circumstances which weakened their probative value. Many cases which hold against the Board involved the discharge of a single employee. See e.g., N.L.R.B. v. Fox Mfg. Co., 5 Cir.1956, 238 F.2d 211; N.L.R.B. v. Birmingham Publishing Co., 5 Cir.1958, 262 F.2d 2; Schwob Mfg. Co. v. N.L.R.B., 5 Cir.1962), 297 F.2d 864, 870; 5 Cir.1962, 297 F.2d 864, 870; 2 Cir.1962, 309 F.2d 898, 905-06. In these cases there was no inference based on the percentage of union adherents discharged to rebut the employer's testimony. Some cases of group firings completely lack any basis for an inference resting on the percentage of union men fired. Thus, in N.L.R.B. v. Atlanta Coca-Cola Bottling Co., 5 Cir.1961, 293 F.2d 300, we held that the Board was not supported by substantial evidence because only half of the employees discharged were union adherents, and half the total number of employees were union adherents; there was no evidence at all that the company knew about the union activities of many of the men discharged. In other cases of group firings, the percentage inference may be overcome by strong evidence that there was 'cause' for the discharge or that the company did not know who the union adherents were. N.L.R.B. v. The Newton Co., 5 Cir., 236 F.2d 438, held that the discharge of a high percentage of union adherents was not discriminatory in that the selection followed a non-discriminatory company rating system. N.L.R.B. v. Moore Dry Kiln Co., 5 Cir. 1963, 320 F.2d 30, 33, held that the inference based on statistics was inadequate, because there was a lack of evidence that the company knew the men were union adherents and because most of the men were discharged on a basis of strict seniority.

It would serve no useful purpose to add pages to this opinion by discussing the case of each employee who was discharged. This Court has read and reread the entire record. We find that only McMen and Cox were discharged for cause.

Except as to the reinstatement of McMen and Cox, we grant enforcement of the Board's order.

1 The Board found it 'unnecessary to consider other conduct also alleged as violations of Section 8(a)(1) of the Act as any findings thereon would not affect the scope of the Order.' We have considered such other conduct to the extent it throws light on the employer's motivation in the interrogation and discharge of the employees

2 Bok, The Regulation of Campaign Tactics in Representation Elections under the National Labor Relations Act. 78 Harv.L.Rev. 38, 106 (1964)

3 Bok, n. 2 at 107

4 Bok, n. 2 at 107

5 'The Board is required to determine the significance of particular acts of interrogation in the light of the entire record in the case.' Blus Flash Express, Inc., 1954, 109 NLRB 591

6 This Court has recognized the coercive effect of interrogation in a variety of cases. The test is whether the questioning tends to be coercive, not whether the employees are in fact coerced. N.L.R.B. v. American Mfg. Co., 5 Cir. 1943, 132 F.2d 760. A few isolated instances of questioning may not be coercive if there is no background of company hostility to the union. N.L.R.B. v. Fuchs Baking Co., 5 Cir. 1953, 207 F.2d 737; N.L.R.B. v. Armour & Co., 5 Cir. 1953, 213 F.2d 625. Nor do a few casual questions violate Section 8(a) (1), if there is no evidence of threats of reprisal or promises of reward. N.L.R.B. v. Mississippi Products Co., 5 Cir. 1954, 213 F.2d 670. Cf. N.L.R.B. v. Harbison-Fischer Mfg. Co., 5 Cir., 1962, 304 F.2d 738, 740 (dissenting opinion). But a few questions may be coercive, if there is even slight evidence of threats or if employees are asked to become informers. N.L.R.B. v. Fox Mfg. Co., 5 Cir. 1956, 238 F.2d 211. Even a single question put to a single employee may be a violation, if there is a background of union hostility. N.L.R.B. v. Griggs Equipment, Inc., 5 Cir. 1962, 307 F.2d 275. As the Court stated in Griggs: 'The statements in and of themselves might possibly be lawful and non-coercive, but viewed in its entirety, there is sufficient evidence of a background of union hostility in context with the statements made to bring them within the proscription of 8(a)(1) of the Act.' 307 F.2d at 278. Questioning a large number of employees is coercive, if there is evidence of union hostility, such as statements by the company president to one or two workers. Martin Sprocket & Gear Co. v. N.L.R.B., 5 Cir. 1964, 329 F.2d 417; accord, N.L.R.B. v. Harbison-Fischer Mfg. Co., 5 Cir. 1962, 304 F.2d 738. Other cases hold that a poll of all or most employees violates Section 8(a)(1) even if there are few other circumstances which show coercive effect. Texas Industries, Inc. v. N.L.R.B., 5 Cir. 1964, 336 F.2d 128; N.L.R.B. v. Aycock, 5 Cir. 1964, 328 F.2d 314. Questioning is also much more likely to be coercive, if there is no attempt to reassure the employees against reprisals. Martin Sprocket & Gear Co. v. N.L.R.B., 5 Cir. 1964, 329 F.2d 417, 419-420. Interrogating key employees during an organizational drive makes the company's anti-union message easier to read than interrogating a few employees at random. See e.g., N.L.R.B. v. Syracuse Color Press, 2 Cir. 1954, 209 F.2d 596, 597, cert. denied, 1954, 347 U.S. 966, 74 S.Ct. 777, 98 L.Ed. 1108, discussed in N.L.R.B. v. Harbison-Fischer Mfg. Co., 5 Cir. 1962, 304 F.2d 738, 744 (dissenting opinion)

7 'We do not require that the examiner's finding be given more weight than in reason and in the light of judicial experience they deserve. The 'substantial evidence' standard is not modified in any way when the Board and its examiner disagree. We intend only to recognize the evidence supporting a conclusion may be less substantial when an impartial, experienced examiner who has observed the witnesses and lived with the case has drawn conclusions different from the Board's than when he has reached the same conclusion. The findings of the examiner are to be considered along with the consistency and inherent probability of testimony. The significance of his report, of course, depends largely on the importance of credibility in the particular case. * * *' 340 U.S. at 496, 71 S.Ct. at 469, 95 L.Ed. at 472

8 Foreman Cook admitted that he questioned only four men, Bounds, Locke, Ledbetter and Irby. Four of the eleven men discharged were under his supervision. These four were the same four whom Cook had questioned about their union activities, Bounds, Locke, Ledbetter and Irby. Cook must have known these four men were union adherents and must have recommended their discharge for that reason

The correlation of men questioned and men discharged on Foreman O'Pry's shift is almost as close. O'Pry testified that he questioned seven men, Carrier, Cox, Clepper, McCall, Baggett, Yeager, and Williams. Four men were discharged from O'Pry's shift. These four were Carrier, Cox, Clepper, and McCall. All four of these men discharged from O'Pry's shift were union adherents and they were among the seven whom O'Pry questioned. According to O'Pry, Carrier, Clepper, McCall and Baggett admitted their union activity and Cox, Yeager, and Williams denied it. All seven were union adherents except Williams. But the testimony of the six union adherents whom O'Pry questioned was very different. Yeager and Baggett both testified that they did not tell O'Pry that they had been to the union meeting. Yeager and Baggett were the two union adherents who were not discharged. Of the four union adherents who were discharged, three admitted that they told O'Pry they had attended the union meeting. The fourth, Cox, admitted that he did not tell O'Pry that he had attended the meeting, but someone else told him that O'Pry had said he knew Cox had been at the meeting. From all this the Board could reasonably infer that O'Pry selected for discharge those who admitted their union activity.

Three employees were discharged who were not under Foreman O'Pry or Cook. Of the three, McMen was discharged summarily on Tuesday, February 13. The second, John Hughes, did not testify at the hearing. Neither O'Pry nor Cook was asked whether he had questioned Hughes. The third, Barnett, testified that Foreman O'Pry did ask him whether he had attended the union meeting. O'Pry admitted that he had talked to Barnett but denied that he questioned him about the union activity. Barnett worked under Foreman Alexander on the first shift, and O'Pry was a foreman on the second shift. It is very curious that O'Pry happened to speak to a man who did not work under him, who was on a different shift, and who happened to be one of the eleven men discharged several days later. The Board could reasonably accept Barnett's testimony.

The connection between questioned men and men discharged is too close to be an accident. If we accept Barnett's testimony that he was questioned, we get the following totals. Of the eleven union men questioned, nine were discharged and the other two testified that they did not admit their union activity. Of the eleven union men discharged, nine were questioned, there was no evidence either way about the tenth, and the eleventh (McMen) was a union leader discharged summarily the day the questioning began. Yet only one of the approximately seventy four nonunion employees was questioned. From all this we can only infer that Camco selected those employees, who, to the best of Camco's knowledge, were union adherents.