大西洋奴隸貿(mào)易中奴隸和船員死亡率的決定因素（2）

2023-08-06 10:06 作者:四夕昭宇 0人讀過(guò) | 我要投稿

## 來(lái)自你的消息:

ANALYSIS OF MAJOR CAUSES OF DEATH

Approach

Given the importance of time in an epidemic process it is fortunate that

surgeons recorded information on a daily basis. The unit of observation

selected for a regression analysis of determinants of mortality is daily

information for a particular ship. We chose a logit model in which the depen—

dent variable takes on the value of 1 if there were one or more deaths from a

particular cause.18 Separate regressions are required for slaves and crew on

the grounds that these groups arrived with substantially different immunities

and lived under considerably different conditions. Because there was contact

with shore and provisions were available from shore during loading but not

during the voyage, it is appropriate to analyze mortality during loading sepa.-

rate from mortality during the voyage.

## 來(lái)自ChatAI的消息:

主要死因的分析

方法

鑒于在流行病過(guò)程中時(shí)間的重要性，幸運(yùn)的是，外科醫(yī)生每天記錄信息。我們選擇了一個(gè)特定船只每天信息作為回歸分析死亡決定因素的觀察單位。我們選擇了一個(gè)邏輯模型，其中依賴變量在特定原因有一個(gè)或多個(gè)死亡時(shí)取值為1。對(duì)于奴隸和船員，需要進(jìn)行單獨(dú)的回歸，因?yàn)檫@些群體的免疫力差異顯著，并且生活條件也有很大差異。因?yàn)樵谘b貨時(shí)與岸上有接觸并可獲得來(lái)自岸上的供應(yīng)，而航行期間沒(méi)有供應(yīng)，因此適當(dāng)?shù)貙⒀b貨期間的死亡率與航行期間的死亡率分開(kāi)進(jìn)行分析。

## 來(lái)自你的消息:

The variables available for study of mortality during loading include

region of origin in Africa, slaves per ton on the ship, seas.on of the year,

and time. These variables are motivated below. Experiments with other varia bles are discussed at the end of this section.

Region. Differences in death rates by location could have been caused by

variations in the rate of exposure to pathogens, the number and virulence of

pathogens, and the immunities and other means of resistence by individuals to

disease. With regard to risks of exposure, the nature of activities required

by the crew while slaving may have differed by region. For example, the dis tance that the ship anchored from shore, whether the ship anchored along the

coast or in a river or delta, the amount of time spent by the crew on ship

versus the interior, and sources of food and water may have been important

determinants of mortality. kiong slaves the risks of exposure may have been

influenced by the duration of time from inland capture to loading, the loca tions of forts or prisons relative to sources of disease, the extent of crowd

ing while being held on shore, and sources of food and water. It is virtually

impossible to acquire reliable information about the number and virulence of

pathogens that existed in the late eighteenth century. However, it should be

recognized that differences could have existed as a function of climate,

topography, migration, and the extent of adaptation of pathogens to hosts.

The immunities of individuals were influenced by the degree of adaptation to a

dise?seenviroriment. Thus the extent to which slaves exported frc a particu lar port were acquired in regions remote or otherwise differing in disease

environment of the port of exit may have influenced mortality. Excluding the

few who had been to Africa or the West Indies previously, the crew generally

had little or no resistence to malaria or yellow fever. The general health

and diet of slaves and crew were important particularly for resistence to gas trointestinal diseases; nutrition apparently has little impact on the course

of malaria or yellow fever.19 Therefore the health of slaves at the time of

initial capture, exertions of the journey to the coast, and diet while being

held on the coast or on board ship may have influenced deaths from dysentery.

In conclusion, it should be recognized that the large number of possible

regional determinants of mortality and the lack of information about most of

them will make it difficult to identify specific sources of locational differ ences in mortality that may be found. It should be noted that the influence

of a particular region on mortality could have changed overtime, but that lack

of data restricts study of this issue.

## 來(lái)自ChatAI的消息:

加載中

用來(lái)研究裝載期間死亡率的變量包括出生地的區(qū)域、每噸船上的奴隸數(shù)量、年份的季節(jié)和時(shí)間。下面將對(duì)這些變量進(jìn)行說(shuō)明。本節(jié)還討論了與其他變量的實(shí)驗(yàn)。區(qū)域。造成不同地點(diǎn)死亡率差異的原因可能是受到病原體暴露率、病原體數(shù)量和毒力、個(gè)體對(duì)疾病的免疫力和其他抵抗手段的變化的影響。就暴露風(fēng)險(xiǎn)而言，奴隸販運(yùn)期間船員所需的活動(dòng)性質(zhì)可能因地區(qū)而異。例如，船遠(yuǎn)離海岸的距離，船只是在沿海停泊還是在河流或三角洲停泊，船員在船上與內(nèi)地的停留時(shí)間之比，以及食物和水的來(lái)源可能是決定死亡率的重要因素。對(duì)奴隸來(lái)說(shuō)，暴露風(fēng)險(xiǎn)可能受到從內(nèi)地被捕到裝船的時(shí)間長(zhǎng)短、堡壘或監(jiān)獄與疾病源的關(guān)系、在岸上滯留期間的擁擠情況以及食物和水的來(lái)源的影響。獲取關(guān)于18世紀(jì)晚期存在的病原體數(shù)量和毒力的可靠信息幾乎是不可能的。然而，應(yīng)該認(rèn)識(shí)到，由于氣候、地形、遷移和病原體對(duì)宿主的適應(yīng)程度等因素的作用，差異可能存在。個(gè)體的免疫力受到對(duì)特定疾病環(huán)境的適應(yīng)程度的影響。因此，從特定港口出口的奴隸是否被獲得于遠(yuǎn)隔地區(qū)或在疾病環(huán)境有所不同的地區(qū)可能會(huì)影響死亡率。除了少數(shù)曾經(jīng)去過(guò)非洲或西印度群島的人外，船員通常對(duì)瘧疾或黃熱病幾乎沒(méi)有抵抗力。奴隸和船員的一般健康和飲食對(duì)于抵抗胃腸道疾病尤為重要；營(yíng)養(yǎng)對(duì)瘧疾或黃熱病的進(jìn)展似乎幾乎沒(méi)有影響。因此，奴隸在被俘初期的健康狀況、到達(dá)海岸的旅途的勞累，以及在岸上或船上的飲食可能會(huì)影響因痢疾而死亡的人數(shù)。總之，應(yīng)該認(rèn)識(shí)到，可能影響死亡率的地區(qū)決定因素眾多，而且對(duì)大多數(shù)這些因素缺乏信息，這將使得很難確定可能存在的地區(qū)死亡率差異的具體原因。應(yīng)該注意到，特定地區(qū)對(duì)死亡率的影響可能隨時(shí)間變化，但因缺乏數(shù)據(jù)，對(duì)此問(wèn)題的研究受到限制。

## 來(lái)自你的消息:

Slaves per Ton. Gastrointestinal diseases such as shigellosis are often

highly contagious.2° Transmission is common through the fecal-oral route,

usually by way of contaminated hands or food. Shigellosis does not confer

long—term immunity because reinfections are frequent, and the multiplicity of

strains precludes resistence to all varieties. Assuming that all instances of

gastrointestinal diseases were not acquired by contamination from a central

source such as food or water, the spread of these diseases would have varied

with density. Thus slaves per ton may have influenced slave death rates. To

the extent that the crew avoided slaves and slave living areas or had resist—

ence, death rates of the crew would have been unrelated to slaves per ton.

Season. The number of sites available for mosquito larvae, and therefore

the mosquito population, may have increased during the rainy season.21 In

addition, mosquitoes have may been more active and survived for longer periods

in humid as opposed to dry weather. Thus the chances of infection from mala ria may have increased during the rainy season. Some observers have noted an

increase in the incidence of 'fevers" during the rainy season.22 Rain and

high humidity may also have prolonged the survival and increased the probabil ity of transmission of bacteria responsible for gastrointestinal diseases.

Time. The spread of infectious disease inevitably involves time. Fac tors that may influence the profile of mortality over time include patterns of

activity that affect the chances of exposure to pathogens, the mechanism of

transmission, the incubation period of the disease, and the interval from the

onset of illness to death. Diseases have characteristic incubation periods

and intervals from the first symptoms to death. Given information about

causes of death, the time profile of mortality may be informative about the

mechanism of transmission and the patterns of activity that led to disease.

The results involving slaves and crew are given in Tables 5 and 6. Since

the chances of deaths should have been a function of the number of persons at

risk, the regressions include the number of slaves (or crew) as independent

variables. The values of -2 log A, where A is the likelihood ratio, show that

each regression is highly significant. Significance tests for groups of coefficients are given in the notes to the tables.

## 來(lái)自ChatAI的消息:

每噸奴隸數(shù)量。痢疾等腸胃疾病通常具有高傳染性。2°傳播通常通過(guò)糞口傳播途徑進(jìn)行，通常是通過(guò)被污染的手和食物傳播。痢疾并不具有長(zhǎng)期免疫力，因?yàn)樵俅胃腥臼穷l繁的，并且多樣性的菌株使得對(duì)所有變種產(chǎn)生抵抗力成為不可能。假設(shè)所有的腸胃疾病實(shí)例都不是源自于中央來(lái)源（比如食物或水）的污染，這些疾病的傳播將會(huì)隨著密度而有所不同。因此，每噸奴隸數(shù)量可能會(huì)影響奴隸死亡率。如果船員避免與奴隸和奴隸居住區(qū)接觸或具有抵抗力，船員的死亡率與每噸奴隸數(shù)量無(wú)關(guān)。季節(jié)。在雨季期間，蚊子幼蟲(chóng)的繁殖地點(diǎn)數(shù)量可能會(huì)增加，從而導(dǎo)致蚊子種群增加。此外，濕度較高的天氣中，蚊子可能更加活躍并且存活時(shí)間更長(zhǎng)。因此，在雨季期間感染瘧疾的機(jī)會(huì)可能會(huì)增加。一些觀察者指出，在雨季中發(fā)熱的情況有所增加。22雨水和高濕度也可能延長(zhǎng)細(xì)菌的存活時(shí)間，并增加導(dǎo)致腸胃疾病的細(xì)菌傳播概率。時(shí)間。傳染病的傳播必然涉及時(shí)間?？赡苡绊懰劳雎孰S時(shí)間變化的因素包括影響暴露于病原體機(jī)會(huì)的活動(dòng)模式，傳播機(jī)制，疾病的潛伏期以及發(fā)病到死亡的時(shí)間間隔。疾病具有特定的潛伏期和從出現(xiàn)癥狀到死亡的時(shí)間間隔。根據(jù)死因的信息，死亡率的時(shí)間分布可以揭示傳播機(jī)制和導(dǎo)致疾病的活動(dòng)模式。關(guān)于奴隸和船員的相關(guān)結(jié)果見(jiàn)表5和表6。由于死亡機(jī)會(huì)應(yīng)該與處于風(fēng)險(xiǎn)中的人數(shù)有關(guān)，回歸分析中包括奴隸（或船員）數(shù)量作為獨(dú)立變量。-2 log A的值，其中A是似然比，表明每個(gè)回歸都非常顯著。關(guān)于系數(shù)組的顯著性檢驗(yàn)見(jiàn)表的注解。

## 來(lái)自你的消息:

Among slaves the deaths from fevers and gastrointestinal diseases by

region of origin were systematically higher in the Bight of Biafra, Compared

to other regions, the probability of deaths among slaves boarded in the Bight

of Biafra was 31 precent higher for gastrointestinal diseases and 58 percent

higher for fevers.23 The Bight of Biafra was also deadly for the crew,

especially for gastrointestinal diseases. However, the chances of crew deatris

from fevers were systematically highest on the Windward Coast and systemati cally lowest on the Gold Coast. The highest chances of crew deaths from gas trointestinal diseases occurred in Senegambia.

Crowding, or slaves per ton, was included in the regressions involving

fevers on the grounds that some gastrointestinal diseases may have been mis diagnosed as fevers. Because the relationship between crowding and mortality

may have been nonlinear, the regressions include a polynomial in slaves per

ton. The tests reported in the notes to Table 6 show that crowding had no

systematic influence on crew deaths. The explanation for the lack of associa tion is unclear, but if the diet and health of the crew was relatively good,

it would have imposed a gastric acid barrier to the spread of gastrointestinal

infections. It is also possible that the crew avoided contact and therefore

contamination from slaves, or that the crew was largely immune (i.e. the crew

was the source of slave gastrointestinal infections) to the diseases involved.

The tests reported in the notes to Table 5 establish that slave deaths were

systematically linked to crowding. Figure 1 shows that the chances of gastrointestinal infection implied by the regression increased roughly ten fold as

the ships approached carrying capacity. This pattern is consistent with

higher probabilities of transmitting infection through contact as density

increased. If some gastrointestinal diseases were misdiagnosed as fevers, it

is not surprising that the chances of deaths from "fevers would have been

related to slaves per ton. Figure 1 shows, however, that the chances of

deaths from fevers eventually declined as slaves per ton rose. The reason for

this pattern is unknown, but it is possible that some of the fevers were

caused by mosquitoes and that for reasons of safety and control of the cargo,

the ships anchored further from shore at hiijh defisities.

## 來(lái)自ChatAI的消息:

在奴隸中，來(lái)自比亞弗拉灣地區(qū)的熱病和胃腸疾病的死亡率比其他地區(qū)普遍更高。與其他地區(qū)相比，在比亞弗拉灣登船的奴隸患胃腸疾病的死亡率高出31%，患熱病的死亡率高出58%。比亞弗拉灣對(duì)船員來(lái)說(shuō)也很危險(xiǎn)，尤其是胃腸疾病方面。然而，來(lái)自熱病的船員死亡率在風(fēng)ward海岸上最高，而在黃金海岸上最低。來(lái)自胃腸疾病的船員死亡率最高的地區(qū)是塞內(nèi)加爾加比亞。

在與熱病相關(guān)的回歸分析中，考慮到有些胃腸疾病可能被誤診為熱病，考慮了人口密度（每噸奴隸數(shù)量）。因?yàn)槿丝诿芏扰c死亡率的關(guān)系可能是非線性的，回歸中包括了奴隸每噸數(shù)量的多項(xiàng)式。表6附注中報(bào)告的檢驗(yàn)結(jié)果顯示，擁擠對(duì)船員死亡沒(méi)有系統(tǒng)影響。對(duì)于這種缺乏關(guān)系的解釋還不清楚，但如果船員的飲食和健康相對(duì)良好，那么可能會(huì)形成胃酸屏障，阻止胃腸感染的傳播。也有可能船員避免與奴隸接觸，避免污染，或者船員對(duì)所涉及的疾病有很強(qiáng)的免疫力（即船員是奴隸胃腸感染的來(lái)源）。

表5附注中報(bào)告的檢驗(yàn)結(jié)果表明，奴隸死亡與擁擠情況有系統(tǒng)關(guān)聯(lián)。圖1顯示，隨著船只接近承載能力，通過(guò)回歸推算的胃腸感染機(jī)會(huì)大致增加了十倍。這種模式與密度增加時(shí)通過(guò)接觸傳播感染的概率增加相一致。如果有些胃腸疾病被誤診為熱病，那么死于"熱病"的機(jī)會(huì)與奴隸每噸數(shù)量相關(guān)并不令人意外。然而，圖1顯示，隨著奴隸每噸數(shù)量的增加，死于熱病的機(jī)會(huì)最終下降。這一模式的原因尚不清楚，但有可能一些熱病是由蚊子傳播的，而出于安全和貨物控制的原因，船只在高密度地區(qū)遠(yuǎn)離海岸停泊。

## 來(lái)自你的消息:

The chances of crew deaths from gastrointestinal diseases were not

systematically related to time as measured by a polynomial in day of loading.

The day of loading coefficients involving fevers were significantly different

from 0 at .10; deaths were systematically lower early in the loading process,

possibly because of the delay between infection and mortality for diseases

such as malaria

Slave deaths from gastrointestinal diseases were systematically higher

during the rainy season (June, July, and August north of the equator and

December, January, and February south of the equator).24 Rain apparently had

no influence on the chances of crew deaths from gastrointestinal causes or on

deaths from fevers. Infective mosquitoes may have been sufficiently prevalent

year-round that the chances of infection varied little by season of the year.

It is also possible t:hat infective mosquitoes were more prevalent during the

rainy season and that patterns of activity, such as expeditions to the

interior, were curtailed in ways that offset the greater risk of infection.

Experiments with other variables include a proxy for the price of slaves,

a measure of association between slave and crew deaths, measures of duration

that slaves had been on board, theage and sex ratios of slaves, recent loading of slaves, year of activity, and month of the year. Since bonuses and

other income received by the captain, the surgeon, and the crew varied with

the value of the cargo, there were incentives to improve care as a function of

slave prices. Unfortunately, reliable annual data on slave prices are lacking

or have not yet been assembled for the late eighteenth century. Experiments

using the price of sugar as a proxy for the price of slaves failed to uncover

any systematic relationship with slave mortality. The incentives at low

pricesmay have been sufficient to elicit all the care that they were willing

to exert, or were able to exert given the primitive state of health care knowledge. The reasons in theory why the price of slaves and the price of sugar

may not have been highly correlated also suggest that sugar prices may have

been a poor proxy for slave prices; for example, fluctuations in weather could

have changed the supply and therefore the price of sugar in ways that had

little influence on slave prices.

In an effort to investigate the possible transmission of gastrointestinal

diseases from slaves to the crew and vice versa, dummy variables representing

death of slaves (or crew) from gastrointestinal causes within the previous

week were included as explanatory variables in the regressions for gastrointestinal diseases. The results showed that there was no systematic interaction between slave and crew mortality.

## 來(lái)自ChatAI的消息:

船員因胃腸疾病死亡的機(jī)會(huì)沒(méi)有與裝船當(dāng)天計(jì)為多項(xiàng)式的時(shí)間相關(guān)。涉及發(fā)熱的裝船當(dāng)天系數(shù)在0.10時(shí)與0明顯不同；在裝船過(guò)程的早期，死亡率明顯較低，可能是因?yàn)橄癔懠策@樣的疾病的感染和死亡之間存在延遲。

奴隸因胃腸疾病的死亡率在雨季（北緯赤道以北的6月、7月和8月，南緯赤道以南的12月、1月和2月）期間明顯較高。雨水顯然不會(huì)影響船員因胃腸疾病而死亡的機(jī)會(huì)，亦不會(huì)影響由發(fā)熱引起的死亡。感染性蚊蟲(chóng)可能在全年范圍內(nèi)普遍存在，因此年份的季節(jié)變化對(duì)感染機(jī)會(huì)的影響可能很小。

還有可能的情況是，感染性蚊蟲(chóng)在雨季更普遍，而活動(dòng)模式（如對(duì)內(nèi)陸的探險(xiǎn)）可能被削減，以抵消更大的感染風(fēng)險(xiǎn)。對(duì)其他變量的實(shí)驗(yàn)包括奴隸價(jià)格的代理變量，奴隸和船員死亡之間的關(guān)聯(lián)度，奴隸在船上的停留時(shí)間，奴隸的年齡和性別比例，最近的奴隸裝載情況，活動(dòng)年份和月份。由于船長(zhǎng)、醫(yī)生和船員所得的獎(jiǎng)金和其他收入隨貨物價(jià)值而變化，因此他們有動(dòng)機(jī)根據(jù)奴隸價(jià)格改善護(hù)理。不幸的是，關(guān)于十八世紀(jì)末奴隸價(jià)格的可靠年度數(shù)據(jù)是缺乏或尚未整理的。使用糖價(jià)作為奴隸價(jià)格的代理變量的實(shí)驗(yàn)未能發(fā)現(xiàn)與奴隸死亡率之間的任何系統(tǒng)關(guān)系。在低價(jià)格環(huán)境下的激勵(lì)可能足以促使他們盡力提供所有可能的護(hù)理，或者鑒于衛(wèi)生保健知識(shí)的原始狀態(tài)，他們只能提供有限的護(hù)理。奴隸價(jià)格和糖價(jià)可能不高度相關(guān)的理論原因也表明，糖價(jià)可能是奴隸價(jià)格的低效代理；例如，天氣的波動(dòng)可能會(huì)改變糖的供應(yīng)，從而對(duì)奴隸價(jià)格影響不大。

為了調(diào)查奴隸與船員之間可能的胃腸疾病傳播情況，將奴隸（或船員）在前一周內(nèi)因胃腸疾病死亡的虛擬變量作為解釋變量放入胃腸疾病的回歸模型中。結(jié)果顯示奴隸和船員死亡之間沒(méi)有系統(tǒng)的相互作用。

## 來(lái)自你的消息:

Because individual slaves were not tagged by date of entry or exit from

the ship, it is impossible to measure accurately the length of time that the

typical slave had been on board. Since life on the ship was probably debili tating, duration aboard may have influenced the incidence and course of

disease. Experiments with measures of average duration constructed on various

assumptions about flows (such as first on, first off and last on, first off)

were unsatisfactory; the procedures sometimes produced impossible results

(negative numbers of slaves at risk on some days, for example) and in those

cases in which the measures could not be rejected out of hand, there was no

systematic association with mortality.

Differential patterns of exposure, immunity, and resistence to disease

may have affected death rates by age and sex. Children, for example, may have

acquired fewer immunities than adults Patterns in African societies of

demand for labor and ccess to food by age and sex and differential treatment

after enslavement could have led to differential mortality for males and

females. In the subset of observations providing the relevant information,

the shares of the cargo enumerated as females and children were included as

explanatory variables. Age and sex composition had no statistically signifi cant influence on fevers, but the chances of deaths from gastromtestinal

causes were systematically higher among females and children.

The chances that someone harbored or transmitted a disease may have

increased with the number of slaves or crew aboard and with the recent addi tion of slaves to the ship. Experiments with these variables failed to un cover any systematic influences. Quadratic terms in the number of slaves and

number of crew, and the number of slaves loaded within the previous week were

statistically insignificant when included in the regressions for fevers and

for gastrointestinal diseases.

Although the data employed in the paper pertain to only 5 years (1792—

1796), it is possible that captains and surgeons adjusted to the incentives of

Dolben's Act in ways that reduced mortality. A linear term in year of the

activity was included in the regressions, and the coefficient was negative and

statistically significant (t=—3.47) for gastrointestinal diseases but statistically insignificant for fevers. During the voyage the coefficient of the

year was statistically insignificant in both regressions. It is possible that

health was improved by methods that applied while loading but not during the

voyage.

## 來(lái)自ChatAI的消息:

由于船上的個(gè)別奴隸沒(méi)有記錄進(jìn)入或離開(kāi)船只的日期，因此無(wú)法準(zhǔn)確測(cè)量典型奴隸在船上的停留時(shí)間。由于船上的生活可能具有削弱體力的效果，停留時(shí)間可能會(huì)影響疾病的發(fā)生和發(fā)展。使用基于不同假設(shè)關(guān)于流動(dòng)情況的平均停留時(shí)間的實(shí)驗(yàn)是不令人滿意的；這些方法有時(shí)會(huì)產(chǎn)生不可能的結(jié)果（例如某些日子奴隸的風(fēng)險(xiǎn)人數(shù)為負(fù)數(shù)），并且在那些無(wú)法完全否定這些測(cè)量值的情況下，并沒(méi)有找到與死亡率有系統(tǒng)相關(guān)的關(guān)系。

暴露、免疫和對(duì)疾病的抵抗力的差異模式可能會(huì)影響不同年齡段和性別的死亡率。例如，兒童可能獲得的免疫能力較少。非洲社會(huì)中對(duì)勞動(dòng)力需求和年齡、性別以及奴役后的差別對(duì)待，可能導(dǎo)致男性和女性的死亡率不同。在提供相關(guān)信息的觀察子集中，將貨物中列舉為女性和兒童的份額作為解釋變量納入考慮。年齡和性別構(gòu)成對(duì)于發(fā)熱沒(méi)有統(tǒng)計(jì)上顯著的影響，但是女性和兒童患胃腸疾病死亡的機(jī)率明顯更高。

某人患病或傳播疾病的機(jī)會(huì)可能會(huì)隨著船上奴隸或船員的數(shù)量增加以及最近添加的奴隸數(shù)量的增加而增加。對(duì)這些變量的實(shí)驗(yàn)沒(méi)有發(fā)現(xiàn)任何系統(tǒng)性的影響。在發(fā)熱和胃腸疾病回歸中包含的奴隸數(shù)量和船員數(shù)量的二次項(xiàng)以及在上一周內(nèi)裝載的奴隸數(shù)量在統(tǒng)計(jì)上都是無(wú)意義的。

盡管本文所使用的數(shù)據(jù)僅涵蓋了5年（1792年至1796年），但船長(zhǎng)和醫(yī)生可能已經(jīng)調(diào)整了行動(dòng)以降低死亡率。在回歸中，年份的線性項(xiàng)被納入考慮，系數(shù)在胃腸疾病中為負(fù)且在統(tǒng)計(jì)上顯著（t=-3.47），但在發(fā)熱中統(tǒng)計(jì)上無(wú)意義。在航行過(guò)程中，年份的系數(shù)在兩個(gè)回歸中都是無(wú)意義的。可能存在在裝載過(guò)程中但在航行中并不適用的方法來(lái)改善健康。

## 來(lái)自你的消息:

The regression models incorporate seasonal phenomena by including a dummy

variable for the rainy season, but in reality seasonal phenomena influencing

health and mortality were more complex than this simple dichotorr'. Patterns

of rainfall, food supply, insect vectors, and human activities varied through-.

out the year and by region in ways that make it difficult to predict the net

influence of all relevant factors.25 Empirical investigation of these phenoni—

ena among slaves is restricted by the small number of deaths that occurred

during loading. mong the crew there is sufficient information to include

dummies in month of loading as regressors in the equation for fevers on ships

that loaded north of the equator. In these data crew deaths were systemati cally lower during the dry season months of March (t=—2.33) and April

(t=-2. 58). Captai.ns were probably aware of seasonal patterns because days

spent loading peaked in March (10.7 percent of all days spent loading) and

were lowest during the summer months of the rainy season.

The Voyage

Although the variables available for study of mortality during the voyage

are similar to those employed for loading, the interpretations may differ.

Region of origin, for example, ecompasses only those aspects such as general

health or patterns of activity immediately before departure that continued to

have an influence during the voyage. Thus contrasts between regional effects

during loading and the voyage may suggest interpretations of the regional

van ables.

Table 7 shows that the chances of slave deaths from fevers were relatively high among ships departing from Sierra Leone, the Bight of Benin, and

Congo-Angola. Senegambia, Sierra Leone, Congo-Angola, and especially the

Bight of Biafra were major sources of slave deaths from gastrointestinal

diseases. The small sample size limited regional comparisons involving the

crew to the Bight of Biafra versus other regions. Companions of Tables 5-8

show that the Bight of Biafra was deadly for slaves and crew. The negative

coefficient for crew size in gastrointestinal diseases is unexpected. Errors

;n crewsize estimates may have contributed to this result, although it does

not register in this way in other regressions. Perhaps there was specialization within the crew in the case of large ships in ways that concentrated

contamination from gastrointestinal diseases.26

Crowding as measured by slaves per ton had no systematic influence on the

chances of deaths for slaves or crew during the voyage. The equations shown

in Tables 7 and 8 include only a linear term in slaves per ton on the grounds

that the range of the variable was small across ships during the voyage; however, a third degree polynomial in slaves per ton is also statistically insignificant. The lack of association contrasts with the systematic association

found for slaves during loading. Figure 1 shows that the chances of slave

deaths tended to level-off at high densities; thus if the lack of association

at high densities during loading applied to the voyage then the finding may be

an artifact of the limited range of observation during the voyage.

## 來(lái)自ChatAI的消息:

回歸模型通過(guò)包括雨季的虛擬變量來(lái)考慮季節(jié)現(xiàn)象，但實(shí)際上影響健康和死亡的季節(jié)現(xiàn)象比這個(gè)簡(jiǎn)單的二元變量更為復(fù)雜。降水、食物供應(yīng)、昆蟲(chóng)媒介和人類活動(dòng)在全年和不同地區(qū)表現(xiàn)出不同的模式，使得很難預(yù)測(cè)所有相關(guān)因素的凈影響。關(guān)于奴隸身上的這些現(xiàn)象的經(jīng)驗(yàn)研究受到了裝貨期間死亡數(shù)量有限的限制。在船員中，有足夠的信息可以將裝貨月份的虛擬變量納入到北緯度以北船只上發(fā)熱的方程中。在這些數(shù)據(jù)中，船員的死亡數(shù)量在三月（t=—2.33）和四月（t=-2.58）這兩個(gè)旱季月份中普遍較低。船長(zhǎng)可能意識(shí)到季節(jié)模式，因?yàn)檠b貨日數(shù)在三月達(dá)到高峰（占所有裝貨日的10.7%），而在雨季夏季的月份最低。

航行旅程

盡管用于研究航行期間死亡的變量與裝貨期間相似，但其解釋可能有所不同。例如，出發(fā)地區(qū)僅涉及到在離港前立即影響航行期間的諸如整體健康狀況或活動(dòng)模式等方面的特征。因此，裝貨期間和航行期間地區(qū)效應(yīng)之間的對(duì)比可能指出了地區(qū)變量的解釋。

表7顯示，從塞拉利昂、貝寧灣和剛果-安哥拉出發(fā)的船只中，因發(fā)熱導(dǎo)致奴隸死亡的幾率相對(duì)較高。塞內(nèi)加爾比紹、塞拉利昂、剛果-安哥拉和特別是比亞弗拉灣是奴隸死于胃腸疾病的主要來(lái)源。樣本規(guī)模小限制了涉及船員與比亞弗拉灣和其他地區(qū)的地區(qū)比較。表5-8的附表顯示，對(duì)奴隸和船員來(lái)說(shuō)，比亞弗拉灣是致命的。在胃腸疾病中，船員規(guī)模的負(fù)系數(shù)是意外的。船員規(guī)模估計(jì)錯(cuò)誤可能導(dǎo)致了這一結(jié)果，盡管這在其他回歸中沒(méi)有體現(xiàn)出來(lái)。也許在大型船只中，船員內(nèi)部存在特殊化，使得胃腸疾病的污染更加集中。

航行期間，以每噸奴隸數(shù)衡量的擁擠程度對(duì)奴隸和船員的死亡幾率沒(méi)有系統(tǒng)影響。表7和8中顯示的方程僅包括奴隸每噸的線性項(xiàng)，因?yàn)楹叫衅陂g該變量的范圍在船只之間非常小；然而，奴隸每噸的三次多項(xiàng)式也在統(tǒng)計(jì)上沒(méi)有重要影響。與裝貨期間的奴隸有系統(tǒng)關(guān)聯(lián)相反，這種缺乏關(guān)聯(lián)性顯著。圖1顯示，奴隸死亡幾率在高密度時(shí)趨于穩(wěn)定；因此，如果在裝貨期間高密度時(shí)的缺乏關(guān)聯(lián)性也適用于航行期間，則這一發(fā)現(xiàn)可能是觀察范圍有限的結(jié)果。

## 來(lái)自你的消息:

Table 8 shows that crew mortality rates from fevers and gastrointestinal

diseases were negatively related to length of loading, but the coefficients

are marginally significant. If there was a systematic association involving

length of loading and these causes of death, then the interpretation may be

that the level of exposure to disease increased during loading such that those

who survived were more immune to attacks and death during the voyage.

Day of the voyage systematically influenced slave deaths from fevers and

gastrointestinial diseases and crew mortality from fevers.27 The relation ships estimated from the regressions are given in Figure 2. In each case the

estimated probabilities of deaths follow an inverted-U pattern as a function

of time. 4mong slaves the peak occurred about 25 days into the voyage for

fevers and 31 days into the voyage for gastrointestinal diseases. Crew deaths

from fevers peaked about 12 days into the voyage. The prevalent and most

deadly forn ofrnalaria in Africa was P. falciparum.28 This form usua!ly has

an incubation period of 10 to 14 days, but may be longer, and may kill up to

25 percent of nonimmune adults within 2 weeks of a primary attack.29 The days

immediately before a ship's departure usually involved considerable loading of

slaves and interaction with shore. The interaction probably increased the

crew's exposure to infective mosquitoes; this combined with the range of.in—

cubation periods and the delay from first symptoms to death is consistent with

the time profile of crew mortality from fevers. Yellow fever runs its course

within 5 to 7 days, and so it is unlikely that infections received on shore

contributed to the peak observed during the voyage. However, the putrid

fevers diagnosed by surgeons are consistent with yellow fever, and it is pos sible that infected mosquitoes taken aboard incubated the virus in ways that

led to deaths from the disease more than one week into the voyage.

Compared to the distribution for the crew, slave deaths from fevers

peaked about 2 weeks later and were more concentrated in the upper tail. The

distribution resembles that for gastrointestinal diseases. Most slaves had

probably been infected with malaria prior to capture, while traveling to the

coast, or while being held on the coast. It is doubtful that slaves with

fever symptoms were loaded. Furthermore, relapses do not occur in P. falci—

parum.30 These considerations combined with the possibilities for misdiagno—

sis and the sensitivity of deaths from Hfeversu to density during crowding

suggest that a large share of the slave deaths attributed to fevers during

the voyage may have been of gastrointestinal origin.

## 來(lái)自ChatAI的消息:

表8顯示，船員因發(fā)熱和胃腸疾病而死亡的率與裝載時(shí)間呈負(fù)相關(guān)，但系數(shù)僅略有顯著性。如果裝載時(shí)間與這些死因之間存在系統(tǒng)性關(guān)聯(lián)，則可能的解釋是在裝載期間，暴露于疾病的程度增加，因此那些幸存下來(lái)的人在航行期間對(duì)襲擊和死亡更具免疫力。

航程的天數(shù)對(duì)奴隸死于發(fā)熱和胃腸疾病以及船員死于發(fā)熱的影響是有規(guī)律的?；貧w估計(jì)得到的關(guān)系如圖2所示。在每種情況下，死亡的預(yù)計(jì)概率都隨時(shí)間呈倒U形。對(duì)于奴隸，發(fā)熱的高峰出現(xiàn)在航行約25天后，而胃腸疾病的高峰則出現(xiàn)在航行約31天后。船員的發(fā)熱死亡在航行約12天后達(dá)到高峰。在非洲，最常見(jiàn)也是最致命的瘧疾形式是惡性瘧原蟲(chóng)。這種形式通常有10到14天的潛伏期，但可能更長(zhǎng)，并且在初次感染的2周內(nèi)可能導(dǎo)致多達(dá)25%的非免疫成年人死亡。船只出發(fā)前幾天通常涉及大量奴隸的裝載和與岸上的互動(dòng)。這種互動(dòng)可能增加了船員暴露于傳染性蚊子的機(jī)會(huì)；加上潛伏期的范圍和從首次出現(xiàn)癥狀到死亡的延遲，這與船員因發(fā)熱而死亡的時(shí)間分布一致。黃熱病的病程持續(xù)5到7天，因此不太可能是在岸上感染導(dǎo)致了航行期間觀察到的高峰。然而，外科醫(yī)生診斷的腐臭性發(fā)熱與黃熱病相一致，可能是由于攜帶進(jìn)船的感染蚊子在船上潛伏導(dǎo)致了一周以上的病死。

與船員的分布相比，奴隸死于發(fā)熱的高峰約晚2周，并且更集中在上尾部。這種分布與胃腸疾病相似。大多數(shù)奴隸在被俘前、前往海岸途中或在海岸上被感染了瘧疾。懷疑被發(fā)熱癥狀的奴隸不會(huì)裝載上船。此外，惡性瘧原蟲(chóng)不會(huì)出現(xiàn)復(fù)發(fā)?？紤]到這些因素以及誤診可能性和在擁擠條件下死于發(fā)熱的敏感性，可以推斷在航行期間歸因于發(fā)熱的奴隸死亡中，很大一部分可能是胃腸疾病導(dǎo)致的。

## 來(lái)自你的消息:

The time pattern of deaths from gastrointestinal diseases shown in Figure

2 is consistent with diffusion by personal contact and with contamination from

a central source. In the case of personal contact the time pattern of deaths

on individual ships would have resembled an inverted—U, and in the latter case

the incidence of the first death across ships would have followed the

inverted-U pattern, but deaths on a particular ship should have been bunched

together, differing only by individual differences in the incubation period.

While it is impossible to measure the relative importance of these processes

from the available data, information in Tables 9 and 10 suggests that spread

by personal contact was a general tendency. Table 9 shows that the average

interval from the beginnng of the voyage to the first death was 17.1 days,

which is significantly less than the interval of roughly 30 days predicted by

the central source hypothesis. The incubation period for shigellosis ranges

from 1 to 6 days and averages 5 days.31 The fact that the average intervals

between deaths 1 and 2 and between 2 and 3 exceeded the average incubation

period casts further doubt on the central source explanation. In addition,

the size and decline of the intervals is consistent with a diffusion by contact process. Table 10 shows that one—half of the ships with 2 or more deaths

from gastrointestinal diseases had average intervals between deaths that

exceeded the upper range of the incubation period of 6 days. Because 10 out

of 26 ships had average intervals under 5 days, however, one cannot rule out

instances of contamination from a central source. Indeed, instances of more

complex processes such as spread by contact combined with contamination from a

iit.i a i uur.e ai We i I WI LII III LII SLO UI LF1e UdL.

## 來(lái)自ChatAI的消息:

圖2中顯示的胃腸疾病死亡的時(shí)間模式與個(gè)人接觸傳播和中央源污染一致。在個(gè)人接觸的情況下，個(gè)別船只的死亡時(shí)間模式將呈現(xiàn)倒U型，而在后一種情況下，首次死亡的發(fā)生率跨越船只將遵循倒U型模式，但是同一船只的死亡應(yīng)該是集中在一起，只有個(gè)體潛伏期的差異。雖然無(wú)法從現(xiàn)有數(shù)據(jù)中衡量這些過(guò)程的相對(duì)重要性，但表格9和10中的信息表明個(gè)人接觸傳播是一種普遍趨勢(shì)。表格9顯示，從航行開(kāi)始到首次死亡的平均間隔為17.1天，明顯小于中央源假設(shè)預(yù)測(cè)的大約30天的間隔。志賀菌病的潛伏期范圍為1至6天，平均為5天。31并且，死亡1和2之間以及2和3之間的平均間隔超過(guò)平均潛伏期，進(jìn)一步對(duì)中央源解釋提出了懷疑。此外，間隔的大小和下降趨勢(shì)與通過(guò)接觸傳播的擴(kuò)散一致。表格10顯示，有2人以上死于胃腸疾病的船只中，一半的船只間死亡的平均間隔超過(guò)了6天的上限范圍。然而，有10艘船只的平均間隔小于5天，因此不能排除中央源污染的情況。事實(shí)上，還存在著更復(fù)雜的過(guò)程，例如通過(guò)接觸傳播與中央源污染相結(jié)合的情況。