In the period following World War II, Satmar hasidim published the hiddushim of the Hatam Sofer with a few critical lines, which address the period of bein ha-shemashot, obscured. Those lines are included among other zemanim-related material at the end of of the first volume of R. Moshe Sofer’s three volumes of hiddushim on the Babylonian Talmud. In a lecture given some twenty years ago, Prof. S. Z. Leiman displayed the altered sefer. The lines in question are not just regular lines, but ones that R. Moshe Sofer transcribed from his rebbe, R. Natan Adler. Years later, Prof. Marc Shapiro found evidence implicating R. Yoel Teitelbaum in the event, something R. Teitelbaum had denied. When the attempt to censor was originally discovered, R. Teitelbaum ordered all volumes returned and the original language restored.
This essay returns to this topic, which I have written about briefly in a different context. In addition to correcting one minor error, I will compare current zemanim for the observance of Shabbat to what was prevalent in Europe during the 17th through 19th century; provide complete background for the teshuvah from R. Sofer, which concerns a baby born late on Shabbat afternoon and which sheds significant light on the edited lines; and explain what the edited lines by R. Adler likely mean.
To set the stage for our discussion, we begin by reviewing the key opinions regarding the onset of shekiah in Halakhah. While there are numerous references to zemanim in Talmudic literature, two sugyot, one in Shabbat and the other in Pesahim, are the most significant.
The sugyah in Shabbat begins with the Mishnah on 34a referring to a period on Friday night described as safek hasheikha, safek eino hasheikha (an uncertainty about whether or not it is dark and already night), and then proceeds to outline an argument concerning the bein ha-shemashot period. That discussion includes an estimate of the length of the bein ha-shemashot period, which is given as the time to walk three-quarters of a mil. Though the sugyah is focused on Friday night, it is almost universally accepted that Shabbat begins Friday night at the beginning of the bein ha-shemashot period and ends Saturday night at end of the bein ha-shemashot period. The sugyah concludes with a statement of Shmuel that night begins with the appearance of three (medium-sized) stars, an oft-quoted rule for determining the end of Shabbat. Despite other important descriptions of various points in the bein ha-shemashot period, the time estimate for the bein ha-shemashot period and the appearance of three stars have come to dominate the halakhic discussion.
The sugyah in Pesahim 94a provides a time estimate for the period from alot ha-shahar to sunrise, and its matching period after sunset in the evening. The sugyah concludes that the duration of the period between alot ha-shahar and sunrise (and the parallel period from sunset to tzeit ha-kokhavim) is the time needed to walk four milin. If the time needed to walk each mil is the commonly assumed eighteen minutes, this translates into a period of (4*18=) 72 minutes. In contrast, the sugyah in Shabbat, when quantifying the period of bein ha-shemashot, sets its length at the time needed to walk three-quarters of a mil, which would be (¾*18=) 13.5 minutes. Given the significant discrepancy between four milin and .75 of a mil, all commentators attempt to resolve the inconsistency by postulating that the two sugyot are focused on different intervals.
The conceptual approach of Rabbeinu Tam on how to resolve the inconsistency posits that the endpoints of the sugyot in Shabbat and Pesahim coincide, providing the point at which Shabbat or any day of the week concludes. To illustrate, let’s assume that we are dealing with a normalized day, around the spring and fall equinox, when sunrise and sunset are at 6:00 a.m. and 6:00 p.m., respectively. Then, according to Rabbeinu Tam, the end of Shabbat is at 7:12 p.m., seventy-two minutes after sunset at 6:00 p.m. However, Rabbeinu Tam continues, the beginnings of the periods differ significantly. While the seventy-two-minute interval specified by the sugyah in Pesahim begins at sunset as commonly defined, the sugyah in Shabbat refers not to sunset, the point at which the sun descends below the horizon, but to a significantly later point when any residual sunlight still visible on the western horizon is about to disappear. The beginning of that “second” sunset occurs at 6:58.5 p.m., which equals the time it takes to walk (4-0.75=) 3.25 milin after sunset proper, 13.5 minutes before 7:12 p.m. At that time, the period of bein ha-shemashot begins; until that time, the day continues, and on Friday night, work is permitted. As Rabbeinu Tam’s opinion was recorded in the Shulhan Arukh O.H. 261 by both R. Yosef Caro and R. Moshe Isserles, it should not be surprising that it maintained widespread acceptance, with most European communities doing work well after sunset on Friday night until the Second World War.
The Geonim, whose work was largely unknown until the modern era, took exactly the reverse position. In their view, the beginning points of the two sugyot (as opposed to the endpoints) are (almost and according to many commentators exactly) identical. Thus, their period of bein ha-shemashot begins at sunset, 6:00 p.m. Shortly thereafter, at 6:13.5 p.m., Shabbat ends and work is permitted. At 7:12 p.m., after the entire time it takes to walk 4-milin period as referred to by the sugyah in Pesahim, all the stars appear, not just the three medium stars mentioned in Shabbat that indicate that Shabbat has ended. That later point when all the stars appear was not ascribed any halakhic significance by the Geonim.
Practice has at various times and for a variety of reasons softened both positions. No one following the Geonim and the Gaon of Vilna (who argued forcefully for a similar position) ends Shabbat 13.5 minutes after sunset on Saturday night, and I do not know of any community that permitted work during the entire 58.5-minute period that the conceptual view of Rabbeinu Tam considers as Friday. Nevertheless, the disagreement is significant. In many European communities, Jews who followed Rabbeinu Tam worked well after what we refer to as sunset on Friday evening; many Jews living in Israel and the Middle East followed the Geonim and ended Shabbat less than thirty minutes after sunset. At least in terms of halakhic theory, a period of approximately forty-five to fifty-five minutes at both the beginning and end of Shabbat is in dispute.
The approach of the Vilna Gaon differentiates the sugyah in Shabbat, which equates the end of Shabbat to the time when three stars appear, from the sugyah in Pesahim, which defines tzeit ha-kokhavim as the time when all the stars appear. On the basis of this, the Vilna Gaon forcefully negates the position of Rabbeinu Tam, who equated the endpoints of the two sugyot. He argues that the sugyah in Pesahim clearly indicates that the period from alot ha-shahar to sunrise is identical in length to the period from sunset and the appearance of all the stars. All the stars appear when no remaining light from the sun impairs their visibility; alot ha-shahar is coincident with the first rays of light in the morning, when (almost) all stars in position to be seen are still visible. In the morning, as more illumination from the sun is discernible, the number of stars that remain visible decreases; in the evening, the reverse process occurs, and as illumination from the sun disappears completely, all the stars that can possibly be seen are visible. A logical consequence of this clear case of symmetry is that the period between sunset and when light disappears, and the period between when light reappears and sunrise, are equal. However, the approach of Rabbeinu Tam, which equates the tzeit ha-kokhavim of Shabbat and the tzeit ha-kokhavim of Pesahim, faces an “unresolvable” issue of asymmetry. Unlike the sugyah in Pesahim, which is silent on the number of stars visible, the sugyah in Shabbat explicitly maintains that the appearance of three medium stars indicate the end of Shabbat. How can the interval when all stars are visible until sunrise equal the interval between sunset and three stars? This challenge to Rabbeinu Tam has never been fully addressed. However, as we will see, this issue led to various positions that deviated from what Rabbeinu Tam’s position would conceptually require.
Now and Then
Currently, the zemanim delineating Shabbat have become largely standardized. Under normal circumstances, with only rare exceptions:
1. Shabbat begins Friday night at sunset.
2. The end of Shabbat for those following the Vilna Gaon is normally either latitude- and season- dependent or some fixed number of minutes after sunset. In the first view, Shabbat in Jerusalem ends 36-42 minutes after sunset and Shabbat in New York ends 39-51 minutes after sunset; the fixed number of minutes for the second view typically falls within that range.
There are still communities whose practice is an outlier, but they are disappearing. Almost universally, the darkness associated with a depression angle of 8.5 degrees, adjusted based on latitude and season, has been used to determine the number of minutes after sunset that bein ha-shemashot (and therefore Shabbat) ends. That determination was made by R. Yehiel Mikhel Tukatchinsky, when he established the Jerusalem calendar, still in use today over a century later.
3. However, those following Rabbeinu Tam almost always wait an unadjusted 72 minutes regardless of where they might be.
It is important to recognize the extent to which current practice differs from that of previous generations. Going back to the 17th century, while professing adherence to Rabbeinu Tam, Shabbat was often observed for (much) fewer than 72 minutes after sunset on Saturday night. For example, R. Avraham Pimential, an acknowledged expert in zemanim who lived in the 17th century, was the first to explicitly mention the impact of both latitude and season on zemanim. Despite his knowledge that being north of Jerusalem and further from the equator implies a later end to Shabbat, he decided that in Amsterdam, Shabbat ends 48 minutes after sunset according to Rabbeinu Tam. R. Pimential was bothered by how Amsterdam, at a latitude almost 20 degrees further north of Jerusalem, could be ending Shabbat (72-48=) 24 minutes earlier. He attributed the difference to the impact of the altitude of Jerusalem as compared to the Dutch lowlands, vastly overestimating the impact of altitude. Around the fall and spring equinox, Amsterdam, being 20 degrees further from the equator than Jerusalem, should end Shabbat not at 72 but 102 minutes after sunset; the difference in altitude results in at most a five-minute difference in Shabbat’s end.
In Europe, where Rabbeinu Tam’s opinion held sway, practices similar to that of R. Pimential were maintained by both R. Yaakov Lorberbaum and R. Moshe Sofer, among others, resulting in Shabbat ending about 52 minutes after sunset for cities slightly north of Amsterdam.  Despite an earlier point at which Shabbat ends, neither R. Lorberbaum or R. Sofer required that Shabbat begin on Friday night more than 20 minutes prior to the time that it ends. It is undeniable that Shabbat’s beginning was well after sunset, and a much shorter period of bein ha-shemashot was the norm, in contrast to what we observe today.
Given limited acceptance of the Geonim’s position in Europe prior to the 19th century, we have little evidence as to how their views were actually practiced. However, in Baghdad, the Ben Ish Hai reports that Shabbat ended twenty to thirty minutes after sunset, which I assume was consistent with the practice in the Middle East.
Relative to historical practice, current observance of the positions of both Rabbeinu Tam and the Geonim is decidedly stricter. We rarely allow Shabbat violations after sunset on Friday night; even those who still follow Rabbeinu Tam also observe the position of the Vilna Gaon regarding Shabbat’s beginning. Additionally, those following the Vilna Gaon delay the end of Shabbat on Saturday night until a depression angle of 8.5 degrees is achieved. If recorded practice in Europe were to be expressed in terms of depression angles, seven to eight degrees (implying a greater degree of illumination) would be more prevalent. Similarly, those following Rabbeinu Tam normally wait a full seventy-two minutes, with a small number even waiting ninety minutes or using latitude and/or season-based adjustments that would further extend Shabbat.
Hatam Sofer’ s Responsum
With this background, we may approach the responsum of R. Moshe Sofer. Teshuvah 80 in Shu”t Hatam Sofer concerns a baby born on Saturday at a time between sunset and the practiced end of Shabbat; its details shed considerable light on the censored lines. The question sent to R. Sofer, requiring an immediate response, was whether the brit should occur on the following Shabbat or Sunday. Today, with almost universal acceptance of the Gaon’s defined start to Shabbat at sunset, such a question would never be asked, and the circumcision would be scheduled for the following Sunday.
Unlike current practice, R. Sofer’s community of Mattesdorf and the nearby community where the baby was born practiced a version of Rabbeinu Tam’s position not uncommon at that time. On Friday night, Shabbat began well past sunset. On Saturday night around the time of the summer solstice, Shabbat ended with the appearance of a requisite number of stars, 52 minutes after sunset, well before the 72 minutes that currently characterizes the practice of Rabbeinu Tam’s position.
We know from sources external to the teshuvah the times of both sunset (8:03 p.m.) and the end of Shabbat (8:55 p.m.) on the day in question, but we do not know precisely when Mattesdorf or the nearby community began Shabbat. By the nature of the she’eilah, however, it is likely that the community started Shabbat on Friday night prior to 8:30 p.m., the time that the baby was born the next day. If the community started Shabbat after 8:30 p.m., there would be no reason to ask a she’eilah; a brit would occur on the following Shabbat.
The fundamental basis of R. Sofer’s reasoning resulted from subtracting the average time to walk ¾ of a mil from 8:55 p.m. to determine the beginning of bein ha-shemashot, as per the opinion of Rabbeinu Tam. R. Sofer, like many other prominent aharonim, disagreed with the Shulhan Arukh that the normally assumed time to walk a mil is 18 minutes. In this ruling, he used his normally assumed position that the time to walk a mil is 22.5 minutes. Therefore, he subtracted the time it takes to walk ¾ of a mil, seventeen (precisely, ¾*22.5 = 16.85) minutes, from the conclusion of Shabbat at 8:55 p.m., and concluded that bein ha-shemashot began no earlier than 8:37 p.m., seven minutes after the baby was born. As a result, R. Sofer proceeds to pasken that the brit should be performed on the next Shabbat.
The style of argument used by R. Sofer has been used by subsequent poskim and raises significant logical issues. In this case, however, given that the date was close to the summer solstice, one can defend the psak (albeit requiring creativity and with difficulty) even for those following the Geonim with respect to the conclusion of Shabbat. Note that using a currently accepted depression angle of 8.5 degrees for the Geonim’s end of Shabbat, we would end Shabbat at 9:08 p.m., thirteen minutes later than 8:55 p.m., which was observed in Mattesdorf. One can only wonder if R. Sofer would have outlawed our present-day stringency for ending Shabbat with his favorite bon mot, hadash assur min ha-Torah.
The Banned Words of R. Adler
Having outlined R. Sofer’s ruling, we now turn to the quote from R. Adler that was later obscured. The Hiddushei Hatam Sofer at the end of Seder Moed contains some halakhic calendrical information. Included in that information was R. Adler’s psak delineating the period of bein ha-shemashot, without providing context, explanation, or justification. Prior attempts to explain R. Adler’s conceptual basis are entirely unsatisfactory; some appear completely unsustainable. R. Adler is quoted as ruling that the bein ha-shemashot period is either twenty-four or thirty-five minutes, choosing whichever is the greater stringency in the case of a biblical violation, and the greater leniency in the case of only a rabbinic violation. How those precise numbers were conceptually derived has never been addressed satisfactorily. Perhaps they were meant to correspond to practice that may not have had a well-defined conceptual basis.
Those who censored the passage likely assumed a rather unlikely interpretation – Rabbi Adler was ending Shabbat 24 – 35 minutes after sunset. There is a story, perhaps apocryphal, of a post WWII agreement between R. Moshe Feinstein and R. Yoel Teitelbaum that in return for Satmar starting Shabbat at sundown, something they certainly did not abide by in Europe, it was agreed that the end of Shabbat would be minimally 42 minutes after sunset. The Satmar, a Hungarian community like R. Sofer’s, likely felt that the rebbe of R. Sofer could not have paskened so differently. Perhaps they thought the lines were inaccurate; in any case, they were not appropriate for publication.
But what had R. Adler in fact ruled? Avoiding for the moment any attempt at a rationale, there are four options that further delimit what R. Adler was suggesting:
- Options 1 and 2: R. Adler is counting forward from the beginning of the period of bein ha-shemashot, either
- option 1) from sunset, or
- option 2) from some other start to the bein ha-shemashot period either
- 2a) less than approximately 27 minutes after sunset, or
- 2b) greater than approximately 27 minutes after sunset.
- Options 3 and 4: R. Adler is counting back from the end of the bein ha-shemashot period, either
- option 3) from the appearance of three small stars, 50 to 55 minutes after sunset, as practiced in Frankfurt and other cities, or
- option 4) from an invariant 72 minutes after sunset.
Fortunately, all but options 2b) and 4) can be easily excluded.
First, let’s examine the possibility that R. Adler was counting forward from sunset, as those who censored this passage probably thought. If this were the case, he would be ending Shabbat exceedingly early; even 35 minutes is more than 15 minutes earlier than other contemporary poskim. Furthermore, were this even a remote possibility, his student, R. Sofer, would never have allowed a brit for a baby who was born 27 minutes after sunset on Shabbat afternoon to occur on the following Shabbat. This is because a brit wrongfully done on Shabbat would certainly be a biblical violation, and since having a bein ha-shemashot of only 24 minutes would be the greater stringency in this case, the baby would have been born after bein ha-shemashot was over. Option 1) is eliminated.
If R. Adler intended to count forward from some unspecified but known time for the start to bein ha-shemashot, that point would have to be more than 27 minutes after sunset (probably at least 30-32 minutes to account for uncertainty about the precise time of the baby’s birth and clocks in general), eliminating option 2a). Otherwise, the baby would have been born during bein ha-shemashot, thus posing a similar challenge to R. Sofer’s psak.
Option 2b), if considered, would certainly not need to be censored. For example, if the start of bein ha-shemashot was 8:37, then R. Adler’s times, 24 or 35 minutes later, would approximate the standard practice of Rabbeinu Tam, hardly deserving censorship. The times for the beginning and end of bein ha-shemashot would mirror those in option 4) discussed below and remains a very plausible alternative.
Now let’s examine the possibility that R. Adler was counting times backward, not forward. If this were so, undoubtedly R. Adler was counting from one well-known zman to establish another; in this case from the end of the bein ha-shemashot period in order to establish its beginning. If that point were the practiced time for the end of Shabbat, approximately 50 to 55 minutes after sunset, or at most 25 minutes after the baby in teshuvah 80 was born, again one must assume that R. Sofer would at least raise the opinion of R. Adler, who specified that the time producing the greater humrah must be used; in this case that would be 35 (and not 24) minutes, which would prohibit biblical violations after 8:20 p.m.. Option 3) is eliminated.
This leaves only one alternative to consider: R. Adler was counting back from the conventional time to walk 4 milin, 72 minutes after sunset. The stricter point is 35 minutes earlier, such that bein ha-shemashot begins 37 minutes after sunset, and the other is 24 minutes earlier, such that bein ha-shemashot begins 48 minutes after sunset. 37 minutes after sunset coincides with R. Lorberbaum’s position for the beginning of the period of bein ha-shemashot, 37.5 minutes after sunset. The other point, 48 minutes after sunset, is a few minutes earlier than the practiced end of Shabbat in Frankfurt and other communities. Thus, R. Adler’s position aligns rulings that during the bein ha-shemashot period rabbinic violations are allowed, with the few added minutes providing a margin of safety or perhaps accounting for tosefet Shabbat. Even though this suggestion can be challenged, it corresponds with the 18th-19th century practice of many cities in that region and the psak of R. Sofer.
Both 4) and 2b) generate similar time intervals for the bein ha-shemashot period. Given the mode of expression then prevalent, especially regarding the position of Rabbeinu Tam, that the beginning of bein ha-shemashot is classically derived by subtracting from its end, option 4) is slightly more likely than option 2b).
R. Adler was saying something that conformed with general practice of that era and provided a humrah – twenty-four to thirty-five minutes – versus the conceptual start of bein ha-shemashot only 13.5 minutes before Rabbeinu Tam’s end of Shabbat. Worthy of being censored? I think not.
Certainly, since the end of the Second World War, the most common mode of expression for times related to the beginning and end of Shabbat uses minutes after sunset almost exclusively; minutes before Shabbat’s end, which was used frequently in the past, is rarely heard. The current mode of expression and a resulting mode of thinking may have, at least in part, inspired the censorship. Additionally, and relatedly, the impact of the Vilna Gaon’s insistence that Shabbat start at sunset and the acceptance of this ruling in many pre-WWII communities may be another factor for the change in the text of R. Sofer’s hiddushim. In any case, the desire to eradicate mention that the Sabbath ends 24 – 35 minutes after sunset was the likely grounds for its deletion, regardless of why it was assumed.
Perhaps, further contributing to the confusion was a change in the significance of sunset proper. While currently sunset has unmistakeable halakhic prominence, it was not nearly as significant in central Europe where the prior discussion was centered, particularly given adherence to the position of Rabbeinu Tam. Furthermore, despite the Vilna Gaon’s convincing attack on Rabbeinu Tam’s position, sunset as the precise beginning of Shabbat is neither a logical consequence nor necessarily even a precise one. In fact, R. Hayyim Volozhin explicitly rejected sunset as the start of Shabbat, lending credence to the possibility (in my mind a strong probability) that the Vilna Gaon expressed his view le’migdar milta, as an attempt to over-correct for what was a (often very) late beginning to Shabbat practiced by followers of Rabbeinu Tam.
In the Middle East and other western Mediterranean communities, which always followed the position of the Geonim versus Rabbeinu Tam, Shabbat began a small number of minutes after sunset. I will cite just two of numerous sources. First, R. Yosef Kapah, probably based on historical Yemenite custom, claims that Rambam supported a beginning to bein ha-shemashot 15 minutes after sunset. Second, R. Shmuel Salant would rule that a baby born after sunset but before the call of the mugrab (the Arab gabbai who alerts worshippers for the fourth prayer service seven to ten minutes after sunset) has his brit on the same day the following week.
Despite the almost simultaneous attack on Rabbeinu Tam by both the first Lubavitcher Rebbe and the Vilna Gaon in the late 18th century, their impact beyond Lithuania and adjacent areas in Russia was limited. It appears unlikely that both their views were known by either R. Adler or R. Sofer.
Nonetheless, in the modern era, sunset has assumed remarkable halakhic prominence that it did not have throughout the 19th century and certainly not to R. Sofer or his Rebbe.
Editing out what is objectionable has a long history; I am led to wonder how often what was deleted was also misunderstood.
We will probably never know with certainty why Satmar seemed to misunderstand the position of R. Adler and did not think of either subtracting from Shabbat’s end or or adding not to sunset but to a point much later. Perhaps an interval of 24 to 35 minutes seemed atypical given a more accepted interval of the time to walk ¾ of a mil (13.5, 16.85, or 18 minutes) with a short period of tosefet Shabbat of 2-3 minutes added.
Given what was presented, the advice of Rambam, whose views on the importance of sunset are disputed, must be mentioned. Rambam in Hilkhot Shabbat (5:4) refers to the period between sunset and the appearance of three stars as “bein ha-shemashot.” His language, “hu ha-nikrah bein ha-shemashot,” “that is called bein ha-shemashot,” might imply that Rambam is providing practical guidance instead of a precise definition. Despite what was discussed, Rambam’s guidance ought not be ignored.
 I refer to R. Natan Adler versus R. Nosson Adler since he was an early Ashkenazi adopter of havarah sephardit.
 Marc B. Shapiro, Changing the Immutable: How Orthodox Judaism Rewrites History.
 Teshuvah 80 in Shu”t Hatam Sofer.
 Rabbeinu Tam’s position is contained in Tosafot in both Shabbat and Pesahim s.v. Rabbi Yehuda omer.
 Rabbeinu Tam is likely to have been influenced by his central European location, which he assumed the sugyah in Shabbat described as well. There is no evidence that he was aware of the impact of latitude on most zemanim that would lead him to recognize that what he was observing in central Europe differed from what one would observe in the Middle East.
He was also likely influenced by the Yerushalmi in the beginning of Berakhot and other sugyot that define tzeit ha-kokhavim, the term used in Pesahim, as the appearance of three (medium) stars. In the sugyah in Shabbat, the phrase tzeit ha-kokhavim does not occur. However, at the end of the sugyah in Shabbat, Shmuel asserts, without any recorded opposition, that the appearance of three medium stars indicates that night has begun, beginning the next day.
 See R. Hayyim Benish, Ha-Zemanim ba-Halakhah, volume 2, chapter 44, which covers European practice extensively.
 In both Hebrew and English, the words yom and day can apply to either the day of the week or the daytime period. Following the insight of the Vilna Gaon in O.H. 461, the sugyah in Pesahim is discussing the daytime period, while the sugyah in Shabbat is discussing the point of transition between days of the week.
 The Gaon of Vilna in O.Ḥ. 459 clarifies this by adding the word “kol” to tzeit ha-kokhavim. Tzeit (kol) ha-kokhavim in Pesahim is not the appearance of just three stars, but the much-later appearance of all stars that are potentially visible.
 The position of the Geonim and the position of the Gaon of Vilna in this area are often viewed as the same. While the Gaon clearly identified sunset as the beginning of the period of bein ha-shemashot, the opinion of the Geonim is less clear; some assume that their start of the bein ha-shemashot is potentially as many as fifteen minutes after sunset. For purposes of this paper, we can accept the broadly maintained assumption that their positions are the same . However, in principle, we strongly question the correctness of that widely assumed viewpoint.
 Restating the argument differently, the sugyah in Pesahim equates both the morning and evening periods to the time it takes to walk 4 milin. Rabbeinu Tam equates the endpoints in the two sugyot, assuming that both endpoints occur when only 3 stars become visible, the usual meaning of tzeit ha-kokhavim. Given that the sugyah in Pesahim asserts the morning and evening intervals are of identical length, the Gaon proves that the only meaning possible for the endpoint in Pesahim is not just 3 stars, but all potentially visible stars, tzeit kol ha-kokhavim. a phrase the Gaon introduced to the halakhic lexicon.
 This question is fundamental to the Gaon’s attack on the position of Rabbeinu Tam. Interestingly, this challenge, along with the others that the Gaon raises, are based on logic, observation, and science, as opposed to halakhic principles. His insights, particularly the full extent of what they would imply considering current scientific knowledge, have not been studied.
 A thorough description of depression angles is provided in my paper “A Categorization of Errors Encountered in the Study of Zemanim” to appear in Hakirah, Spring 2019 and deals extensively with the impact of variations in latitude and season on various zemanim. Briefly, a depression angle is a measure of how far below the horizon the sun is either before sunrise or after sunset; a larger depression angle means the sun is further below the horizon with less residual light still coming from the sun. For example, a depression angle of 8.5 degrees is widely assumed today to equate to the level of darkness that indicates the end of Shabbat according to the Geonim. Poskim must decide a depression angle equivalent for alot ha-shahar, mi-sheyakir, etc. Given a location’s latitude and a specified date of the year, one can calculate precisely the number of minutes before sunrise or after sunset at which (the level of darkness associated with) a given depression angle is achieved.
 R. Pimential’s expertise in zemanim was recognized by R. Avraham Gombiner. His sefer, Minhat Kohen, was carefully organized and argued. Unfortunately, it included two significant errors, one discussed below and the other on the length of the twilight period in the winter. Both errors haunt us to this very day. In an odd but regrettable way, the persistence of his errors is testament to his monumental impact.
 R. Feinstein in Igrot Moshe O.H. (4:62) ruled that ending Shabbat fifty minutes after sunset in New York satisfies the position of Rabbeinu Tam; like R. Pimential, R. Feinstein referenced the appearance of a requisite number of stars. Both R. Lorberbaum and R. Sofer, who paskened similarly, provided no justification for their psakim.
 It is not clear if either was aware of the effect of latitude. See the section of Derekh ha-Hayyim about hadlakat neirot and bein ha-shemashot by R. Lorberbaum and Shu”t Hatam Sofer, Teshuvah 80 and the commentary on Shabbat 34a by R. Sofer.
 The correct wording of various texts in the writings on zemanim of both R. Lorberbaum and R. Sofer have been subject to debate, a topic that will be sidestepped. Derekh ha-Hayyim was published for more than a century as part of a siddur, rather than as a standalone sefer, and was an authoritative source of practical guidance. When that guidance differed from local practice, emendations were often seen as required. In Vilna, for example, an entire chapter was deleted. Similarly, it is reported that R. Sofer wrote his teshuvot quickly and from memory, and as a result, on occasion, he made a few errors that need to be addressed in order that the teshuvah be understood properly. For example, in Teshuvah 80, R. Sofer writes an hour when he likely meant a mil.
 Despite living years after the Vilna Gaon’s death, it is doubtful that either was aware of his extensive views critical of Rabbeinu Tam’s position on both the start and end of Shabbat.
 See R. Hayyim Benish, Ha-Zemanim ba-Halakhah, volume 2, chapter 45, which covers practice in the Middle East in detail.
 In the modern day, only a very few minutes after sunset would generate a she’eilah about the day for a brit.
 The position of R. Feinstein is an exception; see footnote 14.
 R. Sofer was the Rabbi of Mattesdorf before assuming that position in Pressburg.
 R. Sofer clearly believed that bein ha-shemashot was invariant across both season and latitude, a more than reasonable alternative even according to those who apply variation based on latitude and season to determine most zemanim. The arguments for and against variability of the bein ha-shemashot period by season and/or latitude are beyond the scope of this paper.
 See my paper in the Spring 2019 issue of Hakirah, which questions the logic (however not the psak) in the teshuvah. Prominent poskim of the last 50 years have used the same style of argument as R. Sofer, producing more questionable pesakim.
 He did use the phrase at least once to my knowledge, when opposing a humrah requiring a larger minimum size for an etrog. Although he practiced that humrah personally, he opposed its extension to the entire community.
 Ha-Zemanim ba-Halakhah, vol. 2, chapter 46, footnote 77.
 Several options are mentioned in Ha-Zemanim ba-Halakhah. Some try to equate 24 to 35 minutes in Frankfurt with various intervals for the period of bein ha-shemashot in the Middle East. As noted and explained below, the numbers appear to align with practice.
 Of course, both were in use previously.
 This point will be addressed in a future paper on the Gaon’s impact on the study of two sugyot in zemanim.
 See the addition to Ma’aseh Rav, section 19.
 Translated literally as “to guard the thing,” the Gaon likely abandoned halakhic precision in order to ensure a clear change in behavior.
 R. Shneur Zalman of Liadi’s attack on Rabbeinu Tam’s position, included in any standard siddur published by the Lubavitch movement, discusses people who did not start Shabbat until more than 30 minutes after sunset.
 See the various pesakim quoted in R. Benish, Ha-Zemanim ba-Halakhah, volume 2, chapter 45.
 See his commentary in his edition and commentary of Mishneh Torah: Hilkhot Shabbat 5:4.
 Zemanim ke-Hilkhatam by R. Boorstyn, chapter 2, section 1, footnote 7.
 Marc B. Shapiro, Changing the Immutable: How Orthodox Judaism Rewrites History.
 Whether “hu ha-nikrah” is an approximation referring only to the beginning of the bein ha-shemashot period at sunset, only to its end at the appearance of three stars, or to both its beginning and its end is debatable. In my view, it is likely that both are intended as suggested practice, as opposed to either one being a precise halakhic delimiter.