JIA 79 (1953) 0233-0237

233

REVIEWS
LifeContingencies.
By Professor
C. WALLACE JORDAN, F.S.A.
[Society
ofActuaries
TextBook.Pp.xi+331.Published
by the Society
of
Actuaries,
Chicago,
Illinois,
1952. $8.00.1
MEMBERS of the Institute
may welltakepridein the factthat,
in the past,
thetext-book
on LifeContingencies
which has been recommended for reading
by theSociety
ofActuaries,
and by itaparents
theActuarial
Society
of America
and theAmericanInstitute
of Actuaries,
hasbeen theofficial
text-book
of the
Institute.
The studyofthetheoryofLifeContingencies
is not an end in itself,
but onlya means to an end,and accordingly
any text-book
that is to serve its
purposeproperly
must dealwiththepractical
aswellasthetheoretical
aspects
of thesubject.
The practice
oflife
assurance
in theUnitedStates
and Canada
differs
in many respects
from thatin GreatBritain
and itistherefore
a great
tribute
to the work of George King and E. F. Spurgeon that the Institute of
Actuaries
TextBook,Part II,LifeContingencies
by King,published
in 1887,
and subsequently
Life Contingencies
by Spurgeon,first
published for the
Institute
in 1922,shouldforso longhave servedtheneedsof students
in the
continent
of North America.
Spurgeons
book isnow out of dateand the scopeof thenew official
textbook of the Institute
and Faculty,
Lifeand OtherContingencies,
which, at the
timewhen thisreviewisbeingwritten,
isawaiting
publication,
has been mainly
determined
by the requirements
of thepresentsyllabus
of the Institute.
The
Societyof Actuaries
istherefore
fortunate
in having secured the services of
so ablean authority
as Mr C. WallaceJordan,Jr.,Associate
Professor
of
MathematicsatWilliamsCollege
and a Fellowof the Society,
to writea new
text-book
whichgivesspecial
attention
tothebenefits
commonly offered
by the
life
assurance
companiesoftheUnitedStates
and Canada.
The authorexplains
in hispreface
that he has been mindful of the needs of
thosemany actuarial
studentswho must acquire their knowledge of Life
Contingencies
withoutthe assistance
of an instructor.
Accordingly,
he has
givenfulland painstaking
explanations
which should make it easy for the
studentwho is-approaching
thesubject
forthefirst
timeto followthe demonstrations
of thevarious
formulae.He has alsoappendedcopiousexamples to
all the chapters.
The book isdividedintothreeparts.PartI,consisting
of seven chapters,
dealswith the mortality
tableand with net premiums,office
premiums and
reserves
for single-life
assurances
and annuities.
The American flavouris
specially
evidentin Chapter6, which explains
some of the more important
modifiedreserve
systemsand non-forfeiture
provisions.
The developmentof
thesubject
inthispartofthebook isneatand follows
a logical
order,
although
laws of mortality
and annuities
payablem timesa yearareintroduced
at an
earlier
stagethan in Spurgeons
book.Those portions
of Chapter I in which
theauthorinvestigates
theeffect
ofassuminga simplealgebraic
formulafor lx are
a little
tedious
and one wonderswhat benefit
thestudent
isexpected
to derive
from thevarious
examples,
attheend of thischapter,
which require
an arithmeticalevaluation
of functions
based on such formulae. In the chapter on

234

Reviews

annuities
itisa pitythattheauthorhas followed
theconvention
thattheterm
'immediate'
means 'payable
inarrear',
becausethishasledhim totheawkward
expression
deferred
immediatelifeannuity.
PartII contains
six chapters,
coveringjoint-life
annuities
and assurances,
contingent
assurances
and reversionary
annuities.
Althoughthesemattersaredisposed
of in abouthalfas
many pagesas Spurgeonrequired,
spacehasbeen foundforsome ofthemore
complicated
contingent
assurances
and reversionary
annuities
which rarely,if
ever,occurin practice.
In Part III a chapteron stationary
populations
is
placed,
somewhat illogically,
with two chapters
on multiple-decrement
theory
and combinedtables.
For multiple-decrement
theorytheauthoruseshisown
notation:
thechanteron combinedtables
contains
a useful
explanation
of the
notation
usedin Americafordisabilitv
, insurance.
An appendixtothebook contains
The Commissioners
1941StandardOrdinary
(CSO) Mortality
Table,
with monetaryfunctions
at 2%,
and the Life Table
for U.S. WhiteMales,1939-41.
When thelatter
table
was added,thefactthat
a slightly
different
version
of ithad already
been included
in Chapter I seems
to havebeen overlooked.
Also appended to the book are some finite-difference
formulae,
calculus
theorems,
derivatives
of actuarial
functions,
answers to the
examples,
a shortbibliography
and an index to the notation.
When George King wrotetheInstitute
ofActuaries
TextBook,PartII,he
couldhave had no inkling
of the developments
that were to take place in the
theoryof probability.
He was therefore
quitecontentto write,on the first
page of his book:
Couldwe find1OO,OOO
children
allbornatthesamemoment,and couldwe follow
them throughout
life,
and enter
in a column the numbers who remain at the end of
eachsuccessive
yearuntil
allhavepassed
away,we shouldform the column living,
headedwiththesymbollx;where lx
represents
thenumber who attain
the precise
age x.
This interpretation
of the mortality
table,
so simpleand satisfactory
to the
unsophisticated
studentof 1887,hasan airofunreality
to-day.As Mr Charles
A. Spoerlhas so forcibly
put itin hispaper,LifeInsurance
and the Theory of
Probability
(Proc.
Cent.AssemblyInst.
Actuaries,
II,289),
To one who issteeped
in theintricacies
of themodem developments,
just the
thought
ofGeorgeKingsblandassemblage
oflxs
and dzs
enshrined
at the base of
thewholetheory
oflife
insurance
must be wellnighintolerable.
Itsuggests
everythingthatisinadequate,
outmoded and oversimplified.
This does not mean thatthe mortality
tableitself
is outmoded. It means,
among otherthings,
(orso itseemsto thereviewer)
thata mortality
table
with
a radixof IOO,OOOat age o may be regardedas representing,
not the actual
numbers of lives
surviving
to successive
agesout of 1OO,OOObirths,
but the
numbers who areexpected
to survive
to successive
ages out of 1OO,OOO lives
aged o who are takenat random from an indefinitely
large number of lives
whose mortality
isrepresented
by the table.Similarly,
when mortality
table
functions
areapplied
to an individual
life
or a group of lives,
thelife
or lives
must be regardedas havingbeen takenat random from an indefinitely
large
number of lives
who areassumedto experience,
as a whole,themortality
on
which thetableis based.
With thesethoughtsin mind, the reviewer
turnedeagerly
to the opening
pagesof Prof.Jordans
book in orderto seewhetherhe sharedtheseviews,or
whetherhe had some otherideasas to a modem presentation
of the mortality

Reviews

235

tablewhich would enablethestudent

to take a firm grasp of the fundamental
principles
underlying
the application
of the tableto the solution
of practical
problems.It was extremelydisappointing
therefore
to find the author explaining
a mortality
table
witha radixof 1OO,OOOby thestatement
that
thefigures
inthe lx
columnindicate
thenumberofsurvivors
ateachagex and thedx
figures
indicate
thenumberofdeaths
in the year of age x to x + 1.
Worse still,
havingintroduced
thereader
tothefunction
apo,which he calls
the
survival
function
and whichhe denotesby thesymbols(x),
he says (although
he has defined
thefunction
as a probability) :
Letususethesurvival
function
s(x)
...toanswerthefollowing
question;
How many
lives,
outofa groupof 100,000
births,
willsurvive
to age I? The answer is clearly
l00,000
S(I).
No doubtProf.
Jordanhad some good reasons
forpresenting
theseinadequate,
outmodedand oversimplified
ideasto thestudent,
but itissurprising
that his
reasons
are not explained.
This failure
to getto gripswithfundamental
principles
is,not surprisingly,
evident
againin Chapter15 on multiple-decrement
theory,
where the author
omitstopointoutthathe isassuminginseveral
places
thatthedecrements
are
non-selective,
an assumption
which ishardlyeverfulfilled
in practice.
It is,
however,mildlyrefreshing
to seethe multiple-decrement
tablereferred
to as
a mathematical
model.
In a work ofthisnatureitisonlyto be expected
thattheauthorshouldnod
occasionally.
Perhapstherefore
itisa little
ungracious
to pointout thaton
p. 18 he hasfallen
intothecommon errorofsuggesting
that a group of Iives
can
be subject
to the influence
of a forceof mortality.
On p. 29 he saysthat,
in
a select
tablewith a 3-yearselect
period, l25
represents
thesurvivors
atage 25
ofthe l[20]
lives
insured
atage20 and ofthe l[22]
lives
insured
atage,
22 ;thiswas
themisconception
thatlayattherootof Spragues
famoustheoryof damaged
ives.
There isa curiouslapseon pp. 118 and 119 where the rathertricky
problemofequality
of policy
valuesby two different
tables
is investigated.
In
hisanxiety
toavoidSpurgeons
error
offirst
proving
thata condition
is necessary
and then assumingthatitis alsosuffcient,
the authorblindly
followsMr
Thomas N. E. Greville
(T.S.A.III,
533)in thinking
that the most convenient
way to completethe proofisby induction;
he apparently
fails
to realize
that
when theproblemislimited
to a rangeofagesone hasonlytoreverse
the steps
intheproofby which thenecessary
condition
hasbeen established
in order to
show thatthe condition
is sufficient.
He alsomisleadsthe student in the
sentence
beginning
at the bottom of p. 118, where he omits to point out that
the condition
isnot sufficient
unlessk isdefinedas he defines
it in line 3 of
p. 119.
The indexshows signsofhastypreparation;
a few may be mentioned here.
Itisnot clear
why thepage references
for 'Gross
premiums' and 'Premiums,
grossaregivenas 128ff.and 128f.respectively.
No entryisincludedfor
Temporary annuities
(or Annuities,
temporary),Deferred annuities
(or
Annuities,
deferred),
Net premiums or Premiums payablem timesa year.The
page reference
forReserves,
contingent
insurances
is incorrect.
Notwithstanding
thesevariousshortcomings,
the book, on the whole, is
excellent.
It isobvious
that a great deal of loving labour has gone into the
writingof itand itiswellprinted.
The authorhas the giftof being able to
explain
thingsclearly
and he has sparedno painsto make fulluse of thisgift.

236

Reviews

Students
of the Society
will,
no doubt, find the book of great value in their
preparation
fortheexaminations
and they will have good cause to be grateful
to Prof.Jordanforthe careand thoroughness
with which he has developeda
clearand logical
explanation
of the subject.
Studentsof the Institute
might
wellprofit
by reading
thebook,buttheyshouldbe advised
not to take Chapters
P.F.H.
I and 15 too seriously.

Logarithmetica
Britannica,
beinga standard
table
oflogarithms
totwentydecimal
places.
By A. J. THOMPSON, Ph.D. (Lond.)
[Issued
in 9 partsby the Departmentof Statistics,
University
College,
London.
PartII,theninthand last
parttobe issued,
CambridgeUniversity
Press,
1952.
45s.AllotherParts,
ZIS.]
THIRTY yearsafter
he tookup thetask,
Dr Thompsons major work has been
completedby the publication
of the ninth and last part of Logarithmetica
Britannica.
Each of the ninepartscontains
thelogarithms
to twenty decimal
placesof 10,000numbers,eightpartsbeingissuedat fairly
regular
intervals
from 1924 to 1937.The ninthand lastpart had been completed and would
have been readyforpublication
in 1940, but for the war. With the completion
of the fulldesign it can be seen as a worthy memorial of the tercentenary
of
thepublication
of Briggss
Arithmetica
Logatithmica
in 1624.
The olderreadersof the Journalmay remember an article
on The Tercentenary
ofCommon Logarithms
(J.I.A.
LVI,72) which welcomed the publication
ofthefirst
parttobe issued,
PartIX, containing
thelogarithms
of the numbers
90,000to 1OO,OOO.
The article
included
some information
about Henry Briggs.
One of the features
of the tables
has been the series
of prefaces
to,each part
which have recordedallkindsof material
of historical
and generalinterest
relative
to Briggss
logarithms.
What istheneed fora tableof logarithms
to so many decimalplaces?Most
persons,
such as actuaries,
whose work may entail
some computing will never
need a largenumber of significant
figures.
Calculating
machines will serve
their
purposes,
probably
withmore speedin calculation
thanlogarithmic
tables
with4, 5,6 or 7 decimalplaces.
The authorargueswith cogency that the day
of suchtables
is past.
There are,however,occasions
when calculations
havetobe made withmore
significant
figures
thancan be conveniently
handledon a calculating
machine.
For any computerin thatpredicament
the new tables
will be a boon. Prof.
Pearsonstates
thatin statistical
and computinglaboratories
theoriginal
Briggs
and original
Vega areingreater
demand thanany more contracted
logarithmic
tables.
The following
areinstances
withinactuarial
experience
when extended
logarithmic
tables
may be required :
(a) The calculation
of high powers,e.g.isolated
or check values in compound interest
tables,
(b) Computationsbased on recurrence
relationships-because
significant
figures
may be lostateachstageof the computation,
(c) Computations
basedon differences
or derivatives,
(d) Any calculations
involving
operations
in sequence,
and
(e) Standardtables
forthepreparation
of other tables.
Readersmay be interested
to know thatLogatithmetica
Britannica
was used
for the calculations
of log,K100m
and log,A100in the expansions
of ee and

237
Reviews
the problem being the summation,ofan infinite
series
with terms of
alternating
sign(J.l.A.
LxxvI, 155).
The secondand fourthcentral
differences,
separately
computed, are tabulatedbesidethelogarithms
of all5-figure
numbers.The lay-out
is clear-and is
easy to read.The Introduction
surveys the methods of interpolation
for
reading
betweenthelines
of thetables,
namelyinterpolation
by the method
of factors,
by Everetts
formulaand othermethods of differences,
and by
Lagranges
method.The authorgivestwo tables
to facilitate
the method of
the
a n dl o g
factors:
thefirst
is
o fl o g
log
secondisof antilogarithms
forlogsup to .0000450.In the series
of Tracts
for Computersthereis a Tableof coeficients
of Everetts
central-diffrence
interpolation
formula which was also computed by Dr Thompson.
The Introduction
describes
fullyand clearly
the methods used in the
calculation
ofthetables.
The authorconstructed
an integrating
and differencing
machinewhichconsisted
offoursimplemachinesoftherotary
type,so arranged
in stepsthata number on theproductregister
of one machine could be transferredmechanically
to the setting
levers
of the machine below itand thata
number on thesetting
levers
of one machinecouldbe similarly
transferred
to
theproductregister
of the one above it.
The Logarithmetica
Britannica
isa standard
work ofoutstanding
importance,
a magnificent
conception
magnificently
carried
out. The author has enhanced
thevalueof hiswork as a memorialto Henry Briggsby thehistorical
material
which he haspublished
comprising :
Part I. The will of Henry Briggs.
Part II. Translation
of a memoir on the lifeand work of Henry Briggs
writtenin Latinby Thomas Smith,D.D., and publishedin 1707. List of
errors
in Briggss
Arithmetica
Logarithmica
of 1624.
PartIII.Four letters
to Sam Ward, Masterof SidneyCollege,
and one to
Thomas Lydiatin which BriggssaysI am still
at my logarithmes
and can
netherfinishe
them tomy minde nor lett
them alone.
PartIV. Titlepage of thework in which Henry Briggss
Treatise
on the
North-Westpassage
to the South Sea was published.
PartV. Titlepageof Briggss
Arithmetica
Logarithmica,
1624.
PartVI. LettertoJohn PellatTrinity
College,
Cambridge.
PartVII. Titlepageand five
otherpagesofJohn Napiers
CanonisDescriptio
from the later
edition
published
in 1619,afterhisdeath.The reproductions
were chosento illustrate
therespective
contributions
of Briggs and Napier to
the subject.
PartIX. Titlepage and two otherpagesof Briggss
Logarithmorum
Chilias
Prima, 1617.
It isto be hoped thatthe authorwillbe ableto round offhislaboursby
publishing
the life of Henry Briggs for which he has been collecting
the
R.E.B.
material.
M. E. 0.