Before: FOULKE, Chairman; MONTOYA and WISEMAN, Commissioners.


The issues before the Commission are (1) whether the Administrative Law Judge erred in vacating an alleged violation of a standard requiring support of trenches in "unstable or soft" soil, and (2) whether that alleged violation, and another alleged violation involving an employee working in the trench without a hard hat, should be classified as serious. We find that the judge erred in vacating the trench support violation, and that both violations should be classified as serious.[[1/]]

Trumid Construction Co. was the excavation contractor for a condominium construction project in Ossining, New York. It was responsible for grading, digging trenches and laying pipe. On March 6, 1986, a Trumid employee was fatally injured on the project when the side wall of a trench collapsed. Trumid failed to report the accident to OSHA as required by 29 C.F.R. 1904.8.

OSHA received notice of the fatality when an insurance company contacted it in July 1986.[[2/]] OSHA then investigated and issued a citation alleging several serious violations, including the two at issue herein.[[3/]]


Trumid was cited for noncompliance with the safety standard at 29 C.F.R. 1926.652(b), which provided:

Sides of trenches in unstable or soft material, 5 feet or more in depth, shall be shored, sheeted, braced, sloped, or otherwise supported by means of sufficient strength to protect the employees working within them. See Tables P-1, P-2 . . . .

Trumid disputes the Secretary's allegation that this standard applies to Trumid's excavation. Specifically, the parties disagree over whether the sides of the alleged "trench" were 5 feet or more in depth, and whether the alleged trench was in "unstable or soft material." It is undisputed that the sides of the trench were not shored, sheeted, or braced.

A. Whether the cited trenching standard applied to Trumid's excavation

The dimensions and configuration of the excavation are shown in the following cross-section diagram:

The pictorial exhibit referred to above is not available in this format. Please telephone the Review Commission Public Information Office, 202-606-5398, to request a paper copy; TTY: 202-606-5386; FAX: 202-606-5050; e-mail:


The dimensions given above are not disputed by the parties. To clarify them verbally: the first cut into the soil was 18 to 20 inches deep. That cut was about 11 or 12 feet wide. A second, narrower cut was made down the middle of the first cut. That cut resulted in a cavity about 4 to 4 1/2 feet deeper than the first cut. The second cut was about 3 1/2 feet wide at the bottom and 4 feet wide at the top (the "bench" level).[[4/]] Pipe was laid in the cavity. (Here, the trench already had been dug 700 to 800 feet.)

Former 1926.652(b) applied only to trenches five feet or more in depth. However, the judge concluded that the standard was not applicable to the cited excavation because the trench portion was less than five feet deep, and thus he vacated this item. The judge ruled that the trench's depth should not be measured from the original ground level, but from the bottom of the first cut below ground (the bench level).[[5/]] He stated that original ground level was a "non-existent point" that could not be used in determining the dimensions of the cited trench.

On review, the Secretary argues that under the standard's plain wording and intent, this trench should be measured from the original ground level. Trumid endorses the judge's approach. It contends that to measure the trench's depth from the original ground level would require measuring from an imaginary point in mid-air. Alternatively, it argues that if the area above the benches is considered, the excavation would not be a trench, because it was wider at the top than it was deep.

We reject Trumid's argument that the original ground level must be merely imagined in a trench such as this. The original ground level may be ascertained by stretching a string or tape measure between the tops of the walls -- here, specifically, the tops of the walls created by the backhoe's first cut. The depth of any trench with sloped walls must be measured from a point between the sides. Thus, such measurements are not uncommon.[[6/]] In any event, the standard clearly contemplates measuring the trench's depth from the original ground level, where possible. See 29 C.F.R. 1926.652(c) (overall depth of trench includes area above bench) Table P-1 (top of trench is identified as being the "original ground line"); former 29 C.F.R. 1926.653(n) ("trench" defined as a narrow excavation made "below the surface of the ground").[[7/]] This trench's depth is properly measured from original ground level.

We also reject Trumid's argument that if the portion above the benches is considered part of the excavation, it was not a "trench" under the applicable definition because the excavation would then be wider than it was deep. That definition (former 1926.653(n)) stated:

"Trench" -- A narrow excavation made below the surface of the ground. In general, the depth is greater than the width, but the width of a trench is not greater than 15 feet.

The Commission has held that an excavation may be a trench even if it is wider than it is deep. E.g., Heath & Stich. Inc., supra, n. 7 (trench 12 feet wide and 10 to 11 feet deep). Accord, D. Federico Co. v. OSHRC, 558 F.2d 614 (1st Cir. 1977), aff'g, 3 BNA OSHC 1970, 1975--76 CCH OSHD 20,422 (No. 4395, 1976) (trench 13 feet wide and 6 1/2 feet deep).

Also under Commission precedent, the width of a trench is to be measured at its bottom, so that sloping is not taken into account. Heath & Stich, supra.[[8/]] The width at the bottom of this opening (3 1/2 feet) was less than its depth, regardless of whether the depth is measured from the bench mark or the original ground level.[[9/]]

Trumid neither addresses the cited precedent, nor indicates why in its view it is incorrect. We have reconsidered that precedent and reaffirm it. Trumid was on notice of Heath & Stich and D. Federico, because they predated the alleged violation by many years. Thus, we find that Trumid's excavation was a "trench" and that its sides were more than 5 feet deep.

It also bears noting that the new trenching standard, effective March 5, 1990, states explicitly that: (1) a trench's width is to be measured at the bottom, not the top, and that (2) a trench is "narrow" in that its width is small in relation to its length, not its depth. 54 Fed. Reg. at 45,960 (to be codified at 1926.650(b)).[[10/]] Thus, the new standard is consistent with the former standard in these respects.

B. Whether the trench was dug in "unstable or soft material"

The standard cited here applies only where the ground is composed of "unstable or soft material." See p. 2 supra. The judge found that the soil was unstable because: (1) it was "clay type," (2) "ground water" was present from thawing frost, and (3) the Secretary's soil expert, Charles Shimel, testified that it was unstable.[[11/]] Trumid argues that its employees' eyewitness testimony established that the soil was neither unstable nor soft. It also criticizes Shimel's opinion testimony because it was based in large part on his examination of a soil map and on his visit to the worksite about a year after the accident.

Having reviewed the record, we conclude that the evidence fully supports the judge's finding that the soil was "unstable or soft." Shimel's testimony was detailed and authoritative. In particular, he explained how the photographs in evidence showed that the soil was unstable. Those photographs were taken just after the cave-in. They were stipulated into evidence by the parties, were relied on by witnesses for both sides, and their accuracy was not questioned. Shimel explained that they showed cavities in the lower trench wall, caused by soil sloughing off. He concluded that the soil was unstable on that basis.

To rebut the evidence of instability, Trumid points to the testimony of its site supervisor when the cave-in occurred. That supervisor testified that the soil was "firm," "solid" and "cohesive." Trumid also notes that its witnesses were the only ones to see the trench. OSHA's inspector began his investigation four months after the accident. The trench had been filled in. OSHA's expert did not visit the worksite until a year after the accident.

As indicated above, the reason why OSHA was unable to view the trench first-hand was that Trumid failed to report the accident to the agency as required. Thus, we do not fault OSHA for having only circumstantial evidence of the soil type here.[[12/]]

In any event, the eyewitness testimony of Trumid's employees does not rebut the conclusions that Shimel drew after examining the photographs. None of Trumid's witnesses explained why his expert conclusions based on the photographs were incorrect. In particular, the supervisor's statement that the soil generally appeared "firm," "solid" and "cohesive" does not rebut the expert conclusion that some of it was sloughing off, and that the wall therefore was unstable.

None of Trumid's witnesses indicated that they were expert in the identification of soil types. None denied that soil had sloughed off from the walls. One acknowledged that the soil was wet on top from rain or snow, and that it "was just a little mushy maybe" near the top of the trench. The preponderance of the evidence fuIly supports the judge's finding that the soil was "unstable or soft."

The trench, with its undisputed dimensions, would have complied with OSHA's requirements if the soil had been "hard or compact." 29 C.F.R. 1926.652(c). Trumid offered no reason why the trench would have collapsed had it been dug in stable soil.

Thus, the most plausible explanation for the cave-in is that the soil was unstable.

Shimel had other reasons for his conclusion that the soil was unstable. He consulted a U.S. Department of Agriculture map of the area, and testified that it identified the soil type in the area where the accident occurred as "64 Paxton"--an unstable soil for trenching purposes.[[13/]] Based on the photographs, Shimel also testified that the frozen, wet and thawing condition of the soil contributed to its instability. In addition, Shimel relied on his visit to the worksite exactly one year after the accident. At the time of that visit, the excavation at issue had been closed for almost a year. Nevertheless, Shimel was able to observe the soil revealed by other excavations or backfilling on the project, and he was therefore able to determine that the soil type generally conformed to "64 Paxton" soil.

Although Trumid raises strong objections to the evidence based on the soil map and Shimel's visit,[[14/]] this evidence must be viewed in the context of the entire record. The record evidence in toto is probative of the soil type at the accident site. As Shimel testified without contradiction, the information he used was "the best information that [he] could obtain." Again, this is because Trumid failed to report the fatality as required. In the circumstances, the Secretary should be allowed to establish a violation based on the best evidence available.

The soil map, soil report and worksite visit, about which Shimel testified, constitute substantial evidence that the soil at the accident site more likely than not was unstable. That evidence corroborates his unrebutted, expert conclusions based on the photographs in evidence. As noted above, the standard of proof in our proceedings is the preponderance of the evidence. See n. 5 supra. In our view, the evidence here clearly preponderates in favor of the conclusion that the soil was "unstable or soft," as the judge found.

C. Whether the other elements of a violation have been established

The trench was not sloped or otherwise supported in compliance with the standard's requirements. Table P-1, incorporated by reference in the standard, graphically depicts the required slopes for different types of soils. A slope no steeper than approximately 1:1 a 45 angle -- was required for "average soils." "Unstable or soft material" would require at least as much sloping as "average soils."

Shimel gave the only specific testimony regarding the proper dimensions for the excavation. In his opinion, the trench walls should have been sloped 1 1/2 :1 -- that is, 1 1/2 feet horizontally for every foot vertically -- for stability. That would result in an opening at least 20 feet wide at the top.[[15/]] The actual opening was only 11 to 12 feet wide at the top.

Sloping of either 1:1 or 1 1/2 :1 would have resulted in considerably more soil being removed from the trench walls than the bench configuration used by Trumid. Such sloping would thereby have reduced the cave-in hazard at which the standard was directed. Trumid's method of benching did not provide protection that is equivalent to that required by section 1926.652(b) for soft or unstable soil.[[16/]]

It is undisputed that Trumid employees worked in the inadequately protected trench, and that it was aware of the conditions. Thus, all the elements of a violation have been established: the standard applied to the conditions, its terms were violated, Trumid's employees had access to the hazards, and it had the requisite knowledge of the violation. E.g., North Berry Concrete Corp., 13 BNA OSHC 2055, 2056, 1989 CCH OSHD 28,444, p. 37,643 (No. 86-163, 1989).


The next issue is whether the trench support violation and a hard hat violation, as affirmed by the judge, were serious violations. The applicable provision of the Act is 29 U.S.C. 666(k), which provides:

[A] serious violation shall be deemed to exist in a place of employment if there is a substantial probability that death or serious physical harm could result from a condition which exists,. . . . in such place of employment unless the employer did not, and could not with the exercise of reasonable diligence, know of the presence of the violation.

Trumid had the requisite knowledge of the trench support violation, as found above, and the judge correctly found that it was aware that its employee in the trench had no hard hat. The remaining question is whether there was a "substantial probability" that serious consequences "could result."

We find that that test has been met. Supervisor LaBella testified that Jesus Ayala, the employee killed in the cave-in here, was bending over, apparently to retrieve an object, when the soil collapsed. After the soil collapsed, it covered the bottom of the excavation to a height of 2 1/2 feet. Ayala was not wearing a hard hat. The judge found the hard hat item serious for the following reasons:

The photographs show piles of excavated material along the outside of the trench-- this material consisting of sand, stones and other excavated matter could slide down and strike the men [laying pipes] in the trench causing possible serious head injuries. Actually, part of the trench did collapse and cause the death of one of the workers.

The fatality demonstrates the seriousness of the trench support violation.  See Simplex Time Recorder Co., 12 BNA OSHC 1591, 1597, 1984-85 CCH OSHD 27,456, p. 35,572 (No. 82-12, 1985). Also, we find that there was a significant risk of serious head injury in the trench under the circumstances. Both violations properly are classified as serious. See Communications, Inc., 7 BNA OSHC 1598, 1602, 1979 CCH OSHD 23,759, p. 28,813 (No. 76-1924, 1979) (employees likely would suffer serious physical harm from collapse of trench 6'2" deep, particularly if they were in a bending position).


In assessing penalties, the Commission gives due consideration to the size of the business, the gravity of the violation, the employer's good faith, and its history of previous violations. 29 U.S.C. 666(j). The record does not establish whether Trumid was a large or small employer, or whether it had a history of violations. However, the gravity of the trench support violation is a factor that fully justifies the $480 penalty proposed by the Secretary. Also, Trumid's good faith is not unblemished, because it failed to report the fatality as required. A penalty of $480 will be assessed.

The judge assessed a penalty of $200 for the hard hat violation. The parties do not dispute the appropriateness of that penalty on review. The judge's assessment is affirmed.

Thus, the alleged violation of 1926.652(b) is affirmed. That violation and the violation of 1926.100(a) found by the judge are classified as serious violations, with penalties of $480 and $200, respectively.

Edwin G. Foulke, Jr.

Velma Montoya

Donald G.Wiseman

Dated: September 12, 1990






Docket No. 86-1139

                                   U.S. Department of Labor
                                   Office of the Solicitor
                                   New York, New York 10036

                                        For the Complainant

                                    JAMES HARRINGTON, Esquire
                                    Kissam & Halpan, Esquires

                                        For the Respondent


SOMMER, Judge:

On July 25, 1986, Respondent was issued serious citation no. 1 alleging violations of 29 C.F.R. 1926.100(a), 29 C.F.R. 1926.652(b) and 29 C.F.R. 1926.652(h), with a total penalty proposed of $1,360. An other than serious citation which was issued for an alleged violation of 29 C.F.R. 1904.8 with no penalty proposed was not contested.

A hearing was held in New York, N.Y. All parties were represented by counsel who filed post-hearing briefs. No jurisdictional issues are in dispute, the parties having pleaded sufficient facts to establish the Respondent is subject to the Act and the Commission has jurisdiction of the parties and of the subject matter.


Trumid Construction Company, Inc. is a New York Corporation engaged in the construction business. In March 1986 they were engaged in a pipe laying operation in Ossining, New York. The trench collapsed on one of the respondent's employees resulting in a fatality. The Compliance Officer visited the site of the work activity and accident in July 1986 and issued the citations werein based on his investigation. Since the accident and work activity occurred in March 1986 the Compliance Officer observed none of the alleged violations personally.

Citation No. 1, Item 1 - Alleged violation of 29 C.F.R.


The standard which pertains to head protection for employees provides as follows:

Employees working in areas where there is a possible danger of head injury from impact, or from the falling or flying objects, or from electrical shock and burns, shall be protected by protective helmets.

The standard at issue requires that employees "be protected by protective helmets" where there is "a possible danger of . . . injury from impact, or from falling or flying objects. . ."

The evidence demonstrates that two employees were working in a trench laying pipes without having head protection. This was known by the respondent's superintendent. The photographs show piles of excavated material along the outside of the trench - this material consisting of sand, stones and other excavated matter could slide down and strike the men working in the trench causing possible serious head injuries. Actually, part of the trench did collapse and cause the death of one of the workers. Any dislodging of material could strike men working in the trench. In short, the record supports a finding that the respondent's employees were subject to possible head injury while working in the trench and were without protective helmets. Accordingly, I find that there was a violation of 29 C.F.R. 1926.100(a).

The respondent alleges that the Secretary charged the violation of 1926.100(a) was based on possible injuries to employees working in the trench "close to an operating 225 CAT Back hoe" which was not proven and therefore requires a dismissal of this allegation. This argument is without merit. While the Secretary failed to prove that the "possible injury" could occur due to the "Back Hoe", the total evidence produced substantiates a violation of this section because of the possibility of injury as described aforesaid, and is considered an amendment of the citation and complaint to allege the presence of the other possible causes of injury to the employees working without protective helmets. See Federal Rules of Procedure, section 15(b). It is well settled that administrative proceedings are "liberally construed" and "easily amended". N.L.R.B. v. Fant Milling Co., 360 U.S. 301, 79 S. Ct. 1179, 3 L. Ed. 2d 1243 (1959); Usery v. Marquette Cement Manufacturing Co., 558 F2d 902, 906 (5 OSHC 1793, 1796) (2d Cir. 1977).

As Professor Davis has stated, "the most important characteristic of pleadings in the administrative process is their importance. And experience shows that unimportance of pleadings is a virtue". 1 K. Davis, Administrative Law Treatise a Par. 8.04 at 523 (1958).

The respondent failed to object to the introduction of the evidence relevant to the unpleaded allegations, and it had every opportunity to meet the allegations concerning this violation. In sum, the Secretary's enlargement or amendment of the allegations concerning a violation of this section was proper and the issue not raised by the pleading was tried by implied consent. Considering the circumstances herein, and consistent with the criteria set forth in section 17(j) of the Act, a penalty of $200 is considered appropriate herein.

Citation No. 1 - Item No. 2 - Alleged Violation of 29 C.F.R. 1926.652(b)

The trenching standard at 29 C.F.R. 1926.652(b) provides:

Sides of trenches in unstable or soft material, 5 feet or more in depth, shall be shored, sheeted, braced, sloped, or otherwise supported by means of sufficient strength to protect the employees working within them. See Tables P-1, P-2 (following paragraph (g) of this section).

In order to prove non-compliance with section 1926.652(b) the Secretary must prove by a preponderance of the evidence that the sides of the trench were: (1) in unstable or soft material; (2) five feet or more in depth; and (3) not shored, sheeted, braced, sloped, or otherwise supported by means of sufficient strength to protect the employees working within them.

The three witnesses (Labella, Della-Donna and Brajas) who personally observed the condition and make up of the soil agreed that it was of a clay type. Moreover, the presence of ground water from thawing frost contributed additionally to the instability of the soil. Shimel, the Secretary's witness testified that based on his perusal of aerial photographs of the land topography in the vicinity of the trench plus interpretations made thereon by the U.S. Departments of Interior and Agriculture, the soil in the trench area was unstable; however, this opinion was watered down by his admission that the soil type could change within a short distance (500 feet) (T102). In short, the totality of the credible evidence substantiates that the soil in the trench was of a clay composition with some wetness present. Under 29 C.F.R. 1926.652(b), this type of soil is unstable, and I so find.

However, the Secretary failed to sustain his burden of proving that the sides of the trench were 5 feet or more in depth. Since the trenching violation allegedly occurred in March 1986 and the OHSA investigation took place in July 1986 when the trench was filled in and the land back to normal the Compliance Officer had no personal knowledge of the scene. LaBella, the superintendent testified to varying depths of the trench ranging from 4 feet to 6 feet (T14, 20). Shimel, another witness guessed at the trench depth being "anywhere from eight to ten feet" (T69). Scott, the Compliance Officer testified from what LaBella told him "it was between four and a half to five, maybe five and a half feet to the benchmark". Shimel based his estimate on the depth from the very bottom of the trench to some non-existant point where the ground lie before the cut started which is patently erroneous. The trench line starts from the area where the men are standing (see C-5) and the ground is flat astride the cut, not some non-existant sphere. LaBella testified that the trench "was only filled in a little more than two feet" after the cave in, and a three foot shovel inserted in the trench shovel "six or eight inches of the top". Thus hereto there is a distinct possibility of less than a five foot trench being herein.

In short the evidence as constituted is insufficient to prove this necessary element of the violation. The opinions advanced by the witnesses are speculative and are not persuasive as to the depth of the trench. The varying estimates do not constitute the substantial evidence required to make a finding that the trench sides were 5 feet or more in depth.

The proof must be clear and convincing, not presumed or inferred. Ellison Electric, 1 OSHRC 547, 1 BNA OSHC 3034, 1971-3 CCH OSHD Par. 15,133 (No. 412, 1972); not based on speculation, Edison Lampworkes, 7 BNA OSHC 1818, 1979 CCH OSHD Par. 23,913 (No. 76-484, 1979); or mere conjecture, Franklin Lumber Company, Inc., 74 OSHRC 46 E/3, 2 BNA OSHC 1077, 1079, 1973-4 CCH OSHD Par. 18,206 (No. 900, 1974). Since it was not established that the trench was five feet or more in depth, I find there was no violation of 29 C.F.R. 1926.652(b).

Citation No. 1 - Item No. 3 - Alleged Violation of 29 C.F.R. 1926.652(h)

The standard provides as follows:

When employees are required to be in trenches 4 feet deep or more, an adequate means of exit, such as a ladder or steps shall be provided and located so as to require no more than 25 feet of lateral travel.

The Secretary alleged that the respondent violated 29 C.F.R. 1926.652(h) because it failed to maintain an adequate means of exit from the trench. The undisputed evidence established that the trench was 40-42 feet long and over four feet in depth. There was a ramp or slope in the ground provided which was used by the two employees for entrance and egress which was located 40-42 feet away where the employees were first starting to lay pipe. The employees were laying pipe in the trench, lack of 12 1/2 foot sections. Thusly, it is apparent that while laying the first of the pipe lengths these workmen were over 25 feet away from their point of exit. The evidence of record fully establishes that there was a violation of the standard as alleged. Since an exit had been provided for the employees, though not within the required distance, under the circumstances herein, I find that the violation was non-serious and assess a penalty of $50.


All findings of fact relevant and necessary to a determination of the contested issues have been found specially and appear in the decision. (See Rule 52(a) of the Federal Rules of Civil Procedure). Any proposed findings of fact and conclusions of law that are inconsistent with this decision are denied.


1. Trumid Construction Co. Inc., at all times material to this proceeding, was subject to the requirements of the Act and jurisdiction of the Commission.

2. The record establishes by a preponderance of the evidence that Trumid

a) Committed a serious violation of 29 C.F.R. 1926.100(a)
b) Did not violate 29 C.F.R. 1926.652(b) and said allegation is vacated
c) Committed a non-serious violation of 29 C.F.R. 1926.652(h).

3. The penalty for violation of 29 C.F.R. 1926.100(a) is $200 and the penalty for violation of 29 C.F.R. 1926.652(h) is $50, both of which are appropriate and reasonable.


Based upon the findings of fact, conclusions of law, and the entire record, it is ORDERED

1. A violation of 29 C.F.R. 1926.100(a) is affirmed and a penalty of $200 ASSESSED

2. A violation of 29 C.F.R. 1926.652(b) is vacated

3. A violation of 29 C.F.R. 1926.652(h) is AFFIRMED as non-serious and a penalty of $50 is ASSESSED.


Judge, OSHRC

DATED: July 22, 1987
Washington, D.C.


[[1/]] The cited trench support standard has been superseded by the revised excavation standards, 29 C.F.R. Part 1926, Subpart P, effective March 5, 1990. Excavations, Final Rule, 54 Fed. Reg. 45,894, 45,959 (Oct. 31, 1989); 54 Fed. Reg. 53,055 (Dec. 27, 1989). However, the cited standard was in effect when the alleged violations took place.

[[2/]] Trumid was cited under 1904.8 for not reporting the accident to OSHA within 48 hours, but it did not contest that citation. OSHA's compliance officer, David Scott, testified that OSHA received no notice of the fatality until the insurance company contacted it.

[[3/]] The Commission has the discretion to review the entire judge's decision once it is directed for review. Commission Rule 92(a); Hamilton Die Cast, Inc., 12 BNA OSHC 1797, 1986-87 CCH OSHD 27,576 (No. 83-308, 1986). However, the other issues decided by the judge were not expressly directed for review. Nor do we see any other reason to review the judge's disposition of those issues. As to those issues, the judge's decision is a final order of the Commission, with the precedential value of an unreviewed judge's decision. Leone Construction Co, 3 BNA OSHC 1979, 1981, 1975-76 CCH OSHD 20,387, p. 24,322 (No. 4090, 1976).

[[4/]] A "bench" is a squared-off section of soil below ground, resembling a bench in appearance, such as the areas on each side of the second cut here.

[[5/]] The judge erred in stating that the proof of a violation must be "clear and convincing." The standard of proof in our proceedings is the preponderance of the evidence. E.g., Astra Pharmaceutical Products, Inc., 9 BNA OSHC 2126, 2131, 1981 CCH OSHD 25,578, p. 31,901, (No. 78-6247, 1981), aff'd on other grounds, 681 F.2d 69 (1st Cir. 1982).

The judge stated that the evidence about the trench's depth was speculative, unpersuasive and insubstantial. However, that statement was based on the judge's view that the trench's depth should be measured from the bench level, not the original ground level. He did not disagree with the trench dimensions discussed above, which are essentially undisputed on review.

[[6/]] Trumid also argues that the Secretary's proposed "original ground level" basis of measurement conflicts with her own inspector's actual approach. It claims that Compliance Officer Scott originally estimated the depth of the trench from the bench mark, not the original ground level. However, the evidence does not indicate that Scott knowingly calculated the trench's depth differently than the Secretary now advocates. Thus, Trumid's claim is rejected.

To explain, Scott's estimates were made before the citation was issued, based on the photos and what Trumid's supervisor told him.There is not evidence that Scott knew about the benches when he made those estimates. (For example, the photographs in evidence do not show benches, because the spoil piles were resting on them instead of being set back from the sides of the excavation.) Scott did not address this issue in his testimony. However, Charles Shimel, the Secretary's expert on soil types, testified that he was not aware that benches had been created until Trumid's employees testified, and that he had gotten the impression from Trumid's answers to interrogatories that only one cut had been made.

[[7/]] See also Heath & Stich, Inc., 8 BNA OSHC 1640, 1643, 1980 CCH OSHD 24,580, p. 30,151 (No. 14188, 1980).

[[8/]] If width were measured at the top, many trenches would cease to be trenches once they were properly sloped. Sheesley and Winters Construction Co., 3 BNA OSHC 1340, 1341-42 n.8, 1974-75 CCH OSHD 19,756, p. 23,570 n.8 (No. 6824, 1975) (views of Commissioner Cleary).

[[9/]] Trumid argues that the correct approach to trench measurement was taken by Chief Judge Coffin, who concurred in the First Circuit's opinion in D. Federico. However, his opinion does not support Trumid. Judge Coffin did say that the much greater width (13') than depth (6 1/2') of the cavity there precluded calling it a trench. 558 F.2d at 617. However, that excavation was 13 feet wide at the bottom, because it was not sloped. 3 BNA OSHC at 1970, 1975-76 CCH OSHD at p. 24,377.

[[10/]] The new definition states:

Trench (trench excavation) means a narrow excavation (in relation to its length) made below the surface of the ground. In general, the depth is greater than the width, but the width of a trench (measured at the bottom) is not greater than 15 feet (4.6 m). . . .

[[11/]] Shimel is a civil engineer who specializes in soils and foundation engineering.

[[12/]] Shimel testified without contradiction that the information he used was "the best information that [he] could obtain."

[[13/]] That soil type was described in the Westchester County, N. Y. soil survey interim report of October 1978. (The Ossining, N. Y. worksite was within Westchester County.) The description of "64 Paxton" soil in that report states, "[s]idewalls in excavations are unstable in the surface and subsoil and are subject to caving." Mr. Shimel testified, based on the photos in evidence, that much of the soil was consistent with "64 Paxton" soil, though some was clay-like.

[[14/]] Trumid argues that the soil map was unreliable because Shimel acknowledged that soiI types may change within 100 to 150 feet. One hundred feet represented only 1/10-inch on the soil map. Trumid also criticizes reliance on the soil map because, as noted above, its employees testified from first-hand observation that the soil was "firm," "solid" and cohesive." These are not characteristics of "64 Paxton" as described in the soil survey (see n.13 supra). Trumid criticizes reliance on Shimel's visit to the worksite because the soil he saw was between 400 feet and 1,500 feet away from the accident site. Trumid argues that those observations are not helpful because Shimel acknowledged that he has found changes in soil type within 100 to 150 feet.

[[15/]] We calculated this 20-foot width as follows. Trench sides had to be sloped from the bottom, if sloping was used, under the cited standard. E.g., Edward Kelly and Sons, Inc., 10 BNA OSHC 1340, 1343, 1982 CCH OSHD 25,884, p. 32,386 (No. 76-2802, 1982). Under Table P-1, the approximate angle of repose for "average soil" was 1:1 (a 45 angle), and less steep slopes were required for soils more unstable than average. Here, each side of the excavation should have been sloped 8 1/4 feet horizontally (1 1/2 times the 5 1/2-foot vertical distance). Thus, the total trench width at the top should have been 3 1/2 feet (width of excavation at bottom), plus 16 1/2 feet, for a total of 20 feet.

[[16/]] The trench apparently would have been inadequate under the new excavation standard as well. That standard provides specifications for permissible benching. However, for the soil type involved here, which Shimel testified required sloping of 1 1/2 :1, benches would not be permissible. Appendix A to Subpart P, Figure B-1.3; 54 Fed. Reg. at 45,969.