azimuth is _.k , which is known , and consequently the tangent may likewise be obtained .
Having thus discovered a near value of p , or of the error in our first assumed latitude , we are enabled so far to correct it in our second hypothesis ; and in a much greater degree to reduce the error in our conclusion .
As it will sometimes , though not often , be necessary to have recourse to this expedient two short tables : might be added , in order to facilitate the operation .
The first might contain the log .
cosines to every degree from the i~th to the goth of the quadrant ; and be so contrived , as likewise to exhibit the log .
sine of any arc , as far as the 75th degree , expressed in minutes and seconds of time ; that , by subtracting the log , cosine of the altitude from the log .
sine of the hour-angle , ( the cosine of declination being considered as equal to radius , ) we might obtain the log .
sine of the azimuth .
This should be -made one of the arguments of the second table ; the other being the cosine of the latitude to every five degrees of the first sixty of the quadrant ; and the table should give us the value of 15 xA To hich , multiplied into the minutes contained in the error of time , would determine with sufficient exactness the quantityp .
We might , indeed , with the same facility , compute the fourth table according to the mean value of the cosine of declination , if it could be supposed that such a degree of precision would ever be required .
Perhaps it would be advisable for the mariner , in order to avoid all distinction of cases , to calculate , in every instance , his incremental area with thelat .
' vried ten minutes , instead of one ; and then he would always be secure of a result sufficiently correct for his purpose .
In the case supposed above , if the real were to differ from the assumed latitude 30 minutes , the error in his conclusion would not exceed the fourth part of a second .
We must observe , that if n ( the approximate value of p , thus deduced ) be the minutes by which the first assumed lat. is varied in our second hypothesis , we ought to multiply the correction , after it is taken from the 2d table , by this quantity ; but , as n is always supposed to be a whole number , the additional trouble arising from this process can .
never be an object of the smallest consideration .
The 2d table , however , may be as conveniently applied to the case , where the lat. is varied ten minutes , as where it is varied only one .
The logarithms which form the argument are the same in both cases , except that the index in the former is less by unity than in the latter .
Another species of inaccuracy originates in our conceiving the fluxion of the lat. to vary as Tz , instead of x Tz ; and it may be useful to ascertain how much our conclusion is affected by this circumstance .
We suppose , in fact , that the increments of the lat. corresponding to the finite times gf and ge are to each other as the areas gc and gb drawn into the cosine of the first lat. ; whereas , in strict propriety , they are to each other as the sum of all the elements of these areas drawn into the cosines of the latitudes belonging to each , i. e. ( if ge _ rn X gf ) we consider these incrementsas being in the ratio ofgcxX:gbxAinsteadof gcx gbx Hence 2kh + mX gb.b I. _- .
_gc m*Alc1*2S L ( rad.s i , } 2gg .
2 gb .
_n r-Cr..L ?
.Lrd=i consequently , gd gd __ dr ( d being neglected ) will express Y1 GD GD 7r the part of BG , or BG , ( when they are supposed to be verysmall , ) GD by which bg exceeds bg .
Let this be multiplied into -pD Xand it will equal dr the part of bg itself by -which r.lay-dx r. sy-r it is greater than bg .
But our method of equalizing the areas will take from bg a portion of this excess bearing to the whole the ratio of t : t + ; and , consequently , bg will be made too small br X. rA._ nearly Hence -tr.sy-r 2zSy-r 2,.s ^r will express the minutes contained in the error arising from this cause .
The same conclusion may be deduced from the formula v= 1Sre * XXI the area gb , if the increm ' ent of the r r ?
~~~ time be equal to ge .
For the same reasons , 1= .
the area gb , if the increment of the time be represented by ge .
Hence gb-gb x lie r svr , ~ g'bgb t __ r and g-bg bg bX Tj T7 =%7y X TW4 T= that part of the whole area gb by which it is made to exceed its proper magnitude ; and miA = the minutes contained in the error which we are investigating .
If , for instance , the latitude be 520 12 ' , and the declination ?
0 , both of the same kind , and one of the altitudes be taken at the distance of Aft the other at the distance of e , from the meridian , this error will amount to TI . !
oth part of the whole difference betwixt the real and the assumed latitude .
Hence , if this difference be ten minutes , the error in the approximation will not be more than -7th part of a second ; and , if the difference be one degree , the error will only be six times greater than in the former case .
It will therefore always be inconsiderable , except when the latitude and declination are nearly equal , and on the same side of the equator .
But , if this cause should be likely to produce an effect of some importance , we may prevent it by computing -an incremental area with the least altitude likewise , and the latitude increased or diminished by a minute , and then taking a mean betwixt them both for the magnitude of the area gc .
Thus we shall generally obtain a conclusion sufficiently exact ; but , if we are desirous of rendering it perfectly so in every instance , we must divide the difference betwixt the real and apparent intervals , in such a manner that the areas assigned to the two observations may be to each other in the same ratio as their correspondent incremental areas .
This division will be readily dispatched by making the difference betwixt the logarithms of the two first terms , in the proportion ( taken from the 1st table ) equal to the difference betwixt the logarithms of the two last , which are given when either of the formulae mW 41 um cos. of merid .
alt .
K orA cosof id lt. .
is employed ; but must be found , if the first method of computing the area gc be adopted .
Suppose that a and b , representing the logs .
of the areas gb and gb in the table , appear from inspection to exceed each other nearly as much as the logs .
of the areas gc and gc ; let c and d be the two next terms in succession , and let m and n be the logs .
of the quantities expressing the incremental areas ; then will a ~bc -.-d:a bM nac the quantity to be added to , or subtracted from , a , to make it of the required magnitude .
We have then only to subtract m from this corrected value of a. , to proceed with the remainder to the second table , and take out the corresponding error in the assumed latitude .
This precaution , however , can very seldom be necessary ; and , even when it is deemed advisable to adopt it , the division may be performed with so little regard to exactness as to render the process easy and expeditious .
But we are not to conclude that because this inaccuracy , and also the two first species that were considered , are likely to be introduced , when the latitude and declination are nearly equal and of the same kind , they will therefore unavoidably exist in these circumstances , On thecontrary , they may always be prevented , when the weather is favourable , by making the observations within a smaller distance from the meridian .
We have only to wait till the increase of the sun 's altitude becomes so slow as not to produce a visible separation of the limbs in two or three seconds , and then we may be assured that the azimuth is small , and consequently that none of these errors will be considerable .
I have hitherto supposed that this method is only to be adopted , when the sun , at each observation , is within fifteen degrees of the meridian ; or ( to speak more accurately ) when both the azimuth and the hour-angle are so small that we may consider their tangents as bearing a given ratio to each other ; and , indisputably , these limits should never be transgressed , when it can possibly be avoided ; for we have seen ( page 79th ) that , whatever be the method employed , the smaller the hour-angle , the greater is the exactness with which the lat. is determined .
Sometimes , however , it will be impossible to make both , or perhaps either of our observations within the distance which I have recommended ; but , even in these cases , our rule may be conveniently applied .
It has already been demonstrated that we can never be subject to any material error in consequence of the inequality of the areas gb and gb , except when the zenith-distance of the sun , at his meridian altitude , is very small ; and , for this case , an effectual remedy has been provided .
We need not , therefore , make any farther remarks upon this species of inaccuracy .
But perhaps it will be imagined , that because we still continue to suppose the areas of the ; figura tain6entium " to represent the increments of the latitude , a considerable error will be introduced .
We-can easily prove , however , that in consequence of the increment of the time being so much diminished by increasing the distance from noon , this error will seldom be of moment enough to claim our attention .
Let T be the tangent of the azimuth , and r the tangent of the hour-angle at the observation farthest from noon ; then it follows , from what has been demonstrated before , that the ratio of gb to gc will exceed the ratio of GB to GC , and , consequently , the ratio of gb to gc , ( supposed to be of a proper magnitude , ) by ml .
mr+2 rS L. Hence it is evident that , by diminishing one of these ratios and increasing the other , as we are directed in the present case , till they become equal , we augment the latter by at portion of this difference expressed by t- ?
X and , on this acu t wl b+t re2rS count , it will be made too great by M. rT But the ratio of gb to gc will itself exceed the ratio of GB to GC , its proper value , by Mn .
m*'+ 2.L ; and , by the mez styr thod of equalizing the ratios , it is suffered to retain a portion of this difference expressed by ; and , therefore , it will be Mni ' y+z.rs made too great bym .
mI .
r8.L .
The sum of MI 29 I tr+zrst + zrsT these quantities , m. ++S2 iooooo t+v sr S-r + sir will determine , in minutes of a degree , the amount of the whole error , when the incremental area gc represents one minute of latitude .
But if , in our second hypothesis , the latitude be varied n minutes , instead of one , and p be the correction required , the p29 I tr + zrst above expression becomes P. -.n + ? ?
2 X~~~ I OOOOO t+vsrrr+2 rS r minutes .
Hence , in the lat. of Cambridge , when s~yr the declination is 2 ?
north , and the sun 's distance from the meridian at one observation is , and at the other 450 , p being equal to io , and n equal to unity , the error will amount to little more than half a second .
But , whenever it is judged necessary to guard against this species of error , we need only diminish the value of P in the last expression ; and this will readily be effected by dividing the difference betwixt the real and computed intervals of time , in such a manner that the portions belonging to the two observations may be to each other inversely as the quantities taken out of the fourth table , according to the instructions before delivered .
When the portion belonging to the greatest altitude is thus obtained , we can immediately deduce the correction to be made in the first assumed latitude , and then proceed to calculate the incremental areas .
'Let us suppose , for instance , the difference betwixt the intervals to be q , and the quantities deduced from the 4th table to be a and b ; then will a= the minutes by which a+b the former latitude is to be varied .
If , however , : there be any considerable uncertainty respecting the latitude , I would recommend the following method of obtaining an approximate value of it , before we begin the process for investigating its real magnitude .
This will effectually .
prep deude , in almost every instance , the various errors that have been described .
The increment of the altitude is equal to x Sz , and , therefore , when the -azimuth is so small that its sine varies nearly as the arc , the whole increase of the altitude , whilst the sun is moving to the meridian , will be Az xi = .
LZA the sine of the hour-angle being considered as equal to the arc itself .
Now the first of the two tables which have just been explained , will enable us to find the value of g- , by subtracting the log .
cosine of the altitude from twice the log .
sine of the hour-angle ; and a third table might be added , to furnish us with the whole expression , ( according to the mean value of As , ) one of its arguments being the quantity " 4 .
already determined , and the other the supposed-latitude .
It might , perhaps , be-advisable to add another column to the first-of these tables , contai-ning twice the logarithmic sine of .
the , hour-angle , as this would in some measure abridge the operation .
We should find it more convenient too , if the last table were to give us the complement of the arc whose value is 2 rather than the arc itself ; because it MDccxc3[xe 0 would then only be necessary to subtract the observed altitude from this complement , and we should immediately deduce the zenith distance of the sun , when he had arrived at the meridian .
This being ascertained , we should have no farther difficulty in finding the latitude to be adopted in the subsequent computation .
The exactness of the conclusion which4 is derived from this process , will necessarily depend upon the degree of certainty withW , which the time is given when the observation is made .
An error , in time may arise , both from an irregularity in the going of the clock and a small inaccuracy in estimating the diffiwence-betwixt the longitude of the place where the altitude is now taken , and that where the time was last determined but these causes , it is evident , can seldom be very considerable .
We may generally , I think , be sure of the time within two or three minutes .
The principle upon which this approximation depends will open to us a more compendious way of finding the area g b , and which may always be pursued with advantage , when one of the altitudes can be taken at a small distance from the meridian .
In the higher latitudes,.indeed , the hour-angle may amount to five or six degrees ; but , when the latitude and declination approach towards an equality , and are of the same denomination , this angle must be restrained within narrower limits .
If , instead of deducing the meridian altitude from the altitude observed , according to the directions before given , we consider it as being actually equal to the meridian altitude , and employ the latitude resulting from this hypothesis in computing the hour-angle , this angle must necessarily be found equal to nothing .
We may , therefore , spare ourselves the trouble of performing this part of the operation , and need only calculate the time belonging to the second observation , which is supposed to be made when the sun is at a greater distance from the meridian .
The error in the assumed latitude will , in this case , be equal to the whole increase of the altitude betwixt the time of observation and n1oon ; and from this consideration we may be enabled , in ) any particular instance , to discover at what distance from the meridian this plan may be safely adopted .
We can extend italso to anynumber of observations that happen to be made within the proper limits , by connecting them successively with one on the opposite side of the meridian , and thus determiini ng the areas gb corresponding to each .
laving afterwards calculated an incremental area with any of the assumed latitudes varied one minute , we must subtract its logarithm -from the logarithmic value of each of the areas gb , and wee shall discover , by means of the second table , the correction to be applied to each of the supposed latitudes .
But the method which has just been , explained might always be safely adopted , within the limits prescribed to the former rule , if , in computing the incremental area , we were to employ the latitude obtained by means either of the 3d and 4th , or of the 8d and 5th tables. .
This precaution would entirely prevent the error which might arise -from the inequality of the ratios of GC to GB , and gc to gb , when the difference betwixt the real and the assumed latitude was very considerable .
The error which is occasioned by the disparity of the hour-angles , might be obviated by the same means thatwere practised in the first method .
There is only one objection , : indeed , to this rule 's being exclusively adopted , when either of the altitudes is taken within the distance at first recommended .
It would be necessary , in that case , to extend the table much farther than we are obliged to do , if we adhere to the former rule ; but the labour is so much abridged by ; this plan , that I am dQubtful 0 , whether its advantages might not be more -than sufficient to coMr. .
pensate this inconvenience .
The'table should at least be -carried to such a length aswould enableus-to pro ee6d in this -mranner , whenever an observation was ma-dewhin five orsix degrees of the meridian .
I must not forget to observe , before I conclude the theory , that although I have directed the altitudes to be taken on different sides of the meridian , it is by no means ' requisite that we should invariably adhere to ' thispree-pt We have seen thef reason , indeed , why it is .
expedient , in mbst instances to prefer thismthod , ; as being generally calculated to produce a much greater degree of exactness in the result .
This , however , is not always the casefor , if one of the observations be made beyond the distance originally prescribed , it is of little importance whether the second altitude be taken-on the same side of the meridian , or not .
But it will sometimes be impossible to make the observations on different sides of noon ; and hence it becomes necessary to inquire in what manner the -real latitude may be discovered in these circumstances .
The clock gives us the interval betwixt the observations equal to ae ae ; and by computation we find ag and ag , and thence weV deduce eg -eg , the difference betwixt the errors in time .
Having then assumed , without any regard to accuracy , two portions of time , corresponding to the two observations , whose difference is the same as the difference betwixt the errors before determined , and which are to each other in the inverse ratio of the hour-angles , we must increase or diminish them both equally , till we get the areas in the first table of the same magnitude , and then we may conclude that we have obtained the proper value of each .
The directions which have been given for ' the prevention of errors in the 'ci former case , when the altitudes are taken on different sides of the meridian , are very easily accommodated to the present ; and it would therefore be -superfluous to bestow any farther consideration upon them .
From a review of the inaccuracies to which this method , in particular cases , may be liable , it appears that none of them can ever be of sufficient importance to affect the .
mariner .
If he only computes the time with each of the altitudes and the latitude by account , and an incremental area with the greatest altitude and the former latitude varied ten minutes , the correction will generally be deduced within much less than a second , and , in the most unfavourable circumstances , within a minute , of the truth .
But the astronomer , in every instance , even when the latitude and declination are nearly equal and of the same kind , by adopting the precautions which have been recommended , may be assured of a result perfectly exact .
If , however , he should entertain any doubts upon this point , he might easily compute a second value of the incremental area with the latitude already determined ; and this , it is evident , would necessarily produce a conclusion not less accurate than if it were obtained from the direct method , The most satisfactory way of proving the utility of this rule , will be to suppose a particular latitude and declination ; with these to compute the altitudes , when the sun is at two given distances from the meridian ; and thence to deduce the latitudee by an application of our own principles .
EXAMPLE I. Let the latitude be 5e 27 ' 50 " north , and the decline .
when the first altit .
is taken 2 ?
An,12"1 , and , when the second is taken , 20 211 3 ' ( of the same vdenomination as the latit .
) one observation being made in the morning , when the sun is 9 1 ' 22 " , and the other in the afternoon , when he is Io ?
' ic " , distant from the meridian .
The altitude in the former case will be\~~~~~~~~~ found equal to 370 44 ' 52 " ; and in the latter to 37 ? ?
1I 20 " .
Observation Ist .
Observation 2d .
Lat. 540 I15 ' o " Lat. 54 i6 ' o " Lat. 54.0 15'o0 Log. of a 9,7868838 9,7868838 9,7820217 d 8,6175181 -8617SI8I 8,6146155 I_ 9,9093281 9,9094190 9,909 3281 dl 18,5268462 I8,526937I _ 18,5239436 a in ' dl =2I,2600376 _,2599467 21,2580781 a2= 0,63Qoi88 a 10,6299733 10,6290390 Log. tang .
= I0,6177463 I0,6176983 i0,6i67094 2X log .
tang .
=2 1,2354926 21,2353966 21,2334128 Log. of r= 8,6178915 8,617891S 8,6149839 S= 10,14z7296 I0,142996i _ 10,1427 296 Computedlogr.ofvY= 9,996iI37 9,9962842 z99911323 True _ 9,9983068 9,99333253 Areagb =21931 9,996137 -Area gb 21930 Areagc 1705 Hence , 9 21931 = 12'51 " , and latit.-54 27 ' " .
gc 1705____ If we employ the formula M= msin comnV Sy M putting the area gc , we shall obtain , very nearly , the same result .
Log. of r= 8,61 78915 S 10=,1427296 y= 9,9961139 S 20,1388435 r M3 = 18),4790480 rr = 9,2395240 Log. cosine 9,9933360 2X log .
cos. 19,9867120 Const. log .
3,1015 Log. s 10,1427296 -gc 3,2309416 gb 4,3410188 Difference =,110oo772 , , producing in tab .
2d .
12 ' 52 " Lat. corrected -540 27 ' 502 " But there is no occasion to write down the logs .
of r , s , and y , again : we may add together the logs. .
of s and , y , and subtract their sum from the log , of rm ' at the same time , as they stand in the former operation , and this will render the latter process extremely short .
Let us resume the preceding example , and compute the area gb by the second method , and the area ge by the formula ma _ m3 , vs sy M3 Complement of alt .
52Q z5 ' 8 " Observation ist .
Observation 2d .
Dedlin .
z 22 32 Lat. 540 37 ' 40 " .
Lat. 540 37 ' 4-0g ___ " Lat. 54 37 40 Log. of a= 9,7820217 d 8,6I46155 _ 9,91I3753 am ; dI 212560309 7710 l i=,628oi54 Log. tang .
10,6156259 2x log .
tang .
= 1,23125 Log ' 18,4711836 Log. r 8,6149839 r2 942355918 S7 s= 10,1487823 Log. cosune 9,9934771 Computed 1og..y V 10,0000000 9,99So180 2x log .
cos. = 9,9869542 True 9,99830o6 9,9933253 Coust .
log.3,1015 Areagb i6932 Area gb16927 Log. areagc= 3,237 g'b + gb 33859 Mean value ofgb =16929 Whose log .
( from 4 tab .
ist ) 4229 c= 3,237 Difference I992 producing from tab .
zd 4994 Lat. corrected 54 ? ?
27 ' 51 EXAMPLE if .
Let the real latitude be 1o0 o ' o " and the two observations be made on different sides of noon one when the sun 's distance from the meridian is 5 o ' a " , and the other when his distance is 100 oa o " ; the decline .
in the former case being 7 ' '40 o , and in the latter 7 ? ?
39 ' 3 " , of the same kind as the latitude .
Then will the greatest altitude be &amp; + 832 ' 28 " , and the least 790 50 ' 48 " .
Ob-se4tion1st .
Observation zd .
Lat- .
iO ? ?
1 ' o0 Lat xoO 9 ' o " Lat. 100 ? ?
1 On Log. of a 9,9980259 9,99802759 9-,993-1449 d= 9,1258124 .
9,1258124 _ 9,1249212 -I -= 92467746 9,2460695 9,2467746 am3 dl 21,6254389 2i,62,61440 2,6z214491 a ml d 10,8127194 10,8130720 10,8107245 Log. tang .
10,80751340,8078745 _ 0,8054699 2X log tang .
21,6150268 _i,6157490 21,6109398 Log. of r 9,1297233 3 9,1297233 ___ 9,1288i6o S 9,2536477 9 , j25297200 __ 9,2536477 Cotmputed log .
of vi 9,9983978 -999839-23 _ 9,9934035 True 9,9983442 _ 9,99335I5 Areag b= 536 9,9983978 Areagbzz20 Areagc 55 gb ?
gb O528 .
Hence , correction 9 ' = and lat , 2 ~~~~~~~~~~55 corrected 0oto 626"/ .
The difference , however , betwixt the real and the corrected lat. would only have amounted to 19 " , if we had determined the area gb by the 1st table , instead of taking a mean betwixt the two areas for its proper value .
This error-might have been expected from the near approach of the lat. to the decline .
and ought therefore to have been guarded against .
It proceeds from the three causes of inaccuracy which , I have shown , must necessarily be combined in these circumstances .
The remedies to be applied are well knowi .
We must find an approximate value of p ( page 89 ) by means of the 3d and 4th -tables , and compute an incremental area with the least , as well as with the greatest altitude .
If this plan had been pursued , the latitude would have been ascertained within less than a second of the truth .
Or , if the weather had been favourable , we might have prevented the error as effectually , by making the observations when the hour-angles were much less ' and approached nearer to an equality .
N. B. The log .
sines and tangents of the decline .
and lat. to be employed , in all , the operations should be taken out at the same time .
But , when the first method of computing the incremental area is adopted , we may avail ourselves , with considerable advantage , of the following expedient .
Instead of taking the whole log .
sine and tangent of the new lat. take the increments of each , or the difference betwixt their respective values in the two suppositions ; and find the increment of the log .
tangent corresponding to the increment of the log , secant t This may readily be effected , when the log , tangent is deduced from the log .
secant in the former process , by taking the rate of increase belonging to each , and thence inferring by an easy proportion the proper increment of the log .
, tangent in the latter case .
The difference betwixt this increment and the increment of r is the value of the area gc .
An instance of the method thus improved shall be given in the next example .
dently of the second incremental area , which I have demon .
strated above can never be necessary , except when the zenith distance at noon is very rsmall .
The error of three seconds arises from a different cause , and -might have been entirely excluded with the assistance of the 3d and 4th tables .
We should , in fact , deduce from these tables:an approximation of 22 ' and , if we calculate the area gc with the lat ... ' o " , it will be found equal to 34762 , and consequently 28172 = 48 " 62 , which , 36762 multiplied by 2o , produces 17 ' 49'-',64 for the correction , and the lat. is found =S 27 ' 49",64 .
It would have been still better , however , to have obtained a near value of the lat. in the first instance , -by means of the 3d and 5th tables .
The work might have been dispatched in rather less time , and the conclusioni would have been as rigidly accurate .
It is obvious that , if the time when the first observation was made could have been ascertained within two or three seconds , the area gb might have been immediately found by either of the two methods which have been explained , without the assistance of a second altitude , or of the first table ; and the latitude determined with much greater -facility , and with sufficient exactness .
When the azimuth indeed , as in the present case , does not exceed four or five degrees , an error of a second in time will not produce an error of more than a second in the result ; and , as the azimuth decreases , the error in latitude , arising from a given error in time , will be diminished in the same proportion .
In general , if z belthe error in the assumed time , the error in the corrected latitude will be nearly equal to IT x. Let us then suppose the time to be givenr at the first observation , and let us determitie the are gb by he seco ndmethod , and the area gc by the formula 2 : c. of imed .
cos.9tfdec n..( ) Comp. of alt .
460 28 ' 41 " Decline .
=87 -8 Lat. assumed =6 16 Log. of uv 3,1015 .
=P 9,8604043 _V 'v9-%90199043 Computed log .
of'y =o , o000000 " ,91043 True _= 9,998659 _ 9,9956171 9,762842o Area gb = z13409 Log. of area gb= 4127 Area gc32034452 -gc= 3 , j203 Difference , s4 producing from ta-b .
2d .
8 ' 23"/ .
Lat. corrected = 540 27 ' 521"{ EXA4MPLE IY .
The latitude , which is supposed in the lastexample to be 540 27'so " , is that of Ravensworth , a village about five miles to the north of Richmond , in Yorkshire , where I resided some months in the summer of 1797 .
During that time I neglected no opportunity of taking the sun 's 'merdian altitude ; and , accordingly , the latitude which is there assumed is the mean result of a considerable number of observations .
Moreover , on the 8th of September , I took , with particular care , four altitudes , when the sun was near the meridian ; and with these I will now.calculate the latitude , in order to apply the 1st and 9d -tables , and likewise to exemplify the method of combining several observations together .
The two first altitudes were taken in the morning , one -at 19',4 " , the other at 17 ' 40 " before twelve o'clock , and were respectively 40 ? ?
49 ' 6 " , and 400 , o ' s1 " ; the two last in the afternoon , one at 17 ' 53fIs the other at 19 ' 44 " past twelve o'clock , and were respectively 40 ? ?
49 ' so " , and +e 48 ' 18 " ' .
Now the longitude , found by the eclipses of Jupiter 's satellites , is nearly 10 40 ' west of 'Greenwich ; and hence the decline .
appears to have been , at the tWO first observations , S ? ?
26 ' 25 " , and at the two last , S0 25 ' 47 " ; the interval betwixt the first and second , and betwixt the third and fourth , being so small that we may consider the decline .
as remaining constant during each of these times .
Observation ist .
Observe .
zd .
Observe .
3d .
Observe .4th .
Lat. 540 29 ' Lat. 54 ? ?
29 ' Lat 5 2 .
540 29 ' Log. of a 9,8153538 9,8156096 9,8154610 9,8152367 d= 8,9768+57 8,9768457 8,9760047 8,9760047 = 9.9105959 9,9g05959 9,9105959 9,9105959 r M3 am= 20,9279122 20z9281680 20,9288604 20,9286361 Log. = 18,8300 ; 52 a rM 10,4639 6u 10,4640840 10,4644302 10,4643180 9,4150276 Log. tang .
10,4366767 10,4368z18 10,4372141 10,4370861 Log cosine 9,984.7972 zX log .
tang .
20,8733534 ZZO/ 4zj76 2O,87+I72Z 2XI COS. 19,9695944 Log. ofr 8,9788063 8,9788063 8,9779578 8,9779578 Log. v= 3,1015 S_ 10,1464648 10,1464648 10 , I 464648 10 , 1 464648 S 10 , 1 464648 C0oput .
1g 9,986z5 .
Log..9 .
g. 3 cozzu .
l &lt; ; y 9,99 8 6+. .
9,998.9147 ... .
9,9988508. .
9,9985948 Lo c_3,218 ist observation .
Time from the argument of ist table 8 xq .
" , by clock x9 ' zn_ ' 4th , _ 7Am_ x18 26 19 , 44 Nearlog .
value of gb from ist table 3,296 36 40 39 8 gb 3,276 36 4o 2)6,572 Error in time .
2 28 First log .
value of gb 3,286 ad observation .
Time from the argument of ist table i6 ' 1zl , by clock 17 ' 4Oe 3d- .
__ ____ 1640 1753 32 5z 35 33 Near log .
value of-g b from ist table 3,294 32 52 _ Rug~~by _ 3,264 , Error in time:z 4X+ 2)6,558 Second log .
value of gb 3,279 Mean 3-2-82 Log. value ofg c 3,2 18 Difference io64 producing in ad tab .
i ' io IO Lat. corrected = 540 27 ' 50 ' We now take the same example again , and apply the second method of computing the area gb , and the formula ; in determining the area gc .
" REMARKS .
i. All the altitudes that are taken on the same side of noon , tend only to correct the error which may be supposed to exist in the greatest of these altitudes , and can have no effect in removing any inaccuracy to which the greatest altitude on the other side of the meridian may be subject .
Hence we must take more than one altitude on each side of noon , if we are desirous of obtaining a very exact conclusion .
2 .
When some of the observations are made in the morning , and others in the afternoon , the smaller the hour-angle , in every instance , the more favourable it will be for our purpose .
But , if we cannot procure an altitude on each side of the meridian , we ought to make one observation when the h'ur-angle is as large as possible , and with this all the rest should be separately combined .
We must be cautious , however , not to let the sun be too near the horizon , lest the apparent altitude should be affected by the uncertainty of the refraction .
A If the clock were to furnish us with the true time , we might combine together any two observations made within the proper limits , without applying to the first table , and 'deduce a very exact correction .
Should there even be a Ismall error in the supposed time , , we might still proceed in the same manner , without being liable to any material inaccuracy , provided the difference betwixt the hour-angles was not very considerable .
The error , indeed , occasioned by adopting this method of finding the two areas , and taking a mean betwixt them for the value of g b , may easily be determined in any particular case .
If t and t ' be the respective tangents of the smaller and greater hour-angles , and i their difference ; z the error of the clock in minutes of time ; and Z the whole error in the time computed with the greater altitude ; then will the error in the result be to the whole correction of the latitude ( L ) : t : z itz tL i. , lie &amp; --dx * ?
r it ____ ~~~tz I , .
SL This2 T2 xL::2 7L 7 , tz error will consequently be equal to r. ; and hence it aps pears that , if we had pursued this method in the last example , and there had been an error of a minute in the time given by the clock , there would not have been an error of a single second in the conclusion .
4 .
If the time were determined by equal altitudes , and one of them were to be employed in computing the area g b , it is manifest that we should entirely exclude the error which has just been considered .
It would be necessary , however , in order to correct by the second observation any inaccuracy that may have occurred in reading off at the first , to move the index , and then bring it apparently to the same position again , before we proceeded to take the second altitude .
EXAMPLE V. On the Both of January , 1799 , the following altitudes of the sun 's lower limb were taken in Trinity College , Cambridge ; the height of the barometer being 29,6 inches , and of the thermometer 3i degrees .
This process will require very little explanation .
The first column contains the observed , double altitudes as they were read off from the sextant ; the second contains the corresponding times given by the clock ; the third contains the times from noon determined- , with sufficient exactness by taking half the interval betwixt the first and last observations , ( in which the altitudes are equal , ) and supposing this to be the time elapsed betwixt the first observation and the sungs reaching the meridian .
The fourth column is formed of the log , cosines of the hour-angles taken immediately from the argument of the first table ; and the last column exhibits , the latitudes deduced from every two corresponding altitudes , and is intended to show the agreement betwixt these results .
It is not necessary that we should employ the tables in this operation : for we may take the log .
, cosines : of the hour-angles -from TAYLOR'S Logarithms , after the time is reduced into degrees minutes , and seconds ; and , by dividing the area gb= 7537 , by the area gc203 , ( the natural number belonging to the log .
3,3081 , ) we shall obtain the correction required .
TAB. V. I loO &gt; 320 30 0 450 SOO 550 600 890 890 890 890 89 ? ?
890 89 ? ?
890 89 ? ?
6,5463 59 ' 30 " 59 ' 32 " 59 ' V34"t 59 ' 36 " S9 ' 38 " 594 +0 59'42 " 5 9 ' 44 !
' 59 ' 46"s 6,765 08 ii I4 18 21 -\26 30 6,941 58 3.0 58 35 58 4z 58 46 58 SJ 58 56 _2 8 15 7,066 07 14 22 Z8 35 58 43 58 51 58 59 7 , 163 57 30 57 38 !
57 49 57 57 5 L4 23 -33 45 7,242 0II 24 _ 33 57 42 57 53 4_ I7 30 7,309 5630 56 42 56 58 8 -Ig -32 57 45 57 59 2 7,367 -0 14 32 56 43 56 56 _IO 25 42 0 7,418 55 30 55 46_ 6_ 19 --34 56 S6 25 57 45 746+ ?
-I8 55 40 55 54 I0 -28 56 47 7 -30 7,505 54 30 54 49 14 -39 55 47 7 -28 56 48 I5 7,543 2I 54 485 14 55 45.8 33 5 659 7,578 53 30 53 50 -22 54 4 ?
0 1 14 5 549 i6 45 7,6I o_ -25 53 56 i6 54 38 3 30 55 59 .
29 7,640 52 30 52 57 30 53 5 15 54 42 11 41 IS EXPLANATION OF THE TABLES .
Tab. I. This table contains two arguments .
One is the distance of the sun from the meridian , which is expressed in hours , minutes , and seconds ; and also in terms of the log .
cosine of the hour-angle .
This argument 'is placed in the first column .
The second stands at the top of the table , and consists of the minutes and seconds contained in the error of time .
The table itself exhibits the logarithm of the area gb corresponding to any given error at any particular observation .
But it was necessary to give the log .
value of this area in both directions from the tangent gd , i. e. both when the computed is greater , and when it is less , than the real interval of time .
These two values are included in the same column , but are separated from each other by a point .
The first quantity expresses the log .
area in the latter , the second expresses it in the former case .
Thus , for instance , in the fifth column we have i36.io8 , and we are to understand that the log .
of the area gb is equal to 156 , when the computed time is too small ; and equal to io8 , when the computed time is too great .
The argument at the top is carried to every twelve seconds of time , or every three minutes of a great circle ; and should be continued as far as four or five minutes of time for the first twelve or fourteen degrees from the meridian ; as far as three minutes for the next ten degrees ; two minutes for the next ten degrees ; and so on , ( the time being diminished as the tangent of the hour-angle increases , ) as far as sixty or sixtyfive degrees , beyond which the table need not be extended .
The argument in the first column should be carried to every eight seconds of time for the first twelve or fourteen degrees ; and to every twenty seconds for the restof the table .
Tab. II .
This table consists of two columns .
The first contains the differences betwixt the logs .
of the area gb and gc , when the latit .
is varied one minute in calculating the incremental area .
These differences are negative as far as the fourteenth , because so far the areagc exceeds the areagb .
In the second column are exhibited the errors in the assumed latit .
corresponding to these differences .
Tab. III .
This table is composed of four columns .
The first contains the computed distance in time from noon ; the second the altitude of the sun ; the third the log .
sine of the hour-angle , or the log .
cosine of the altitude , expressed in the first column ; and the fourth twice the logarithm in the former &amp;c. column , being intended to save us troble of multiplying by 2 , when we are finding the logarithm of At * Tab. IV .
In this table there are two arguments .
One of them occupies the first column , and consists of the log .
sines of the azimuth , which are found by subtracting the log , cosine.of the altitude from the log .
sine of the hour-angle , as they stand in the preceding table .
The other argument is placed at the top of the columns , and is formed of the assumed latitude .
The table is designed to exhibit the product of the taangent of the azimuth multiplied into the cosine of the latit .
and into 1'5 ; so that , by taking out this quantity , and multiplying it into the minutes contained in the error of time , we may at once determine how much the assumed latit .
is to be varied in computing the incremental area .
Tab. V. There are two arguments in this table likewise .
That which stands in the first coltumn is the -log .
value of A -obtained from the second table by subtracting the log , cosine of the altitude from twice the log , sine of the hour-.angle .
The second argument , which is placed bove the columns -is the supposed latitude .
The table itself contains the comple.ent of the meridian altitude -at the place of observation ' .
VII .
A Method of inding the Latitude of a. Place , by Means of " two Altitudes of the Sun and the Time elapsed betwixt the Observations .
By the Rev. W. Lax , A. , M. Lowndes 's Profissor of Astronomy in the University of Cambridge .
Read January io , 1799 .
I HOPE the following method of determining the latitude , by means of two altitudes of the sun and the time elapsed betwixt the observations , will be found not less convenient for nautical purposes than the rules which are commonly employed .
But I would rather recommend it in those cases where rigid accuracy is required , and the astronomer is provided with no better instrument for taking the sun 's altitude than a HADLEY'S sextant of the most improved construction .
The process will be neither difficult nor tedious ; and , if the observations are made with proper exactness , I conceive the latitude will generally be ob.taimed within a few seconds of the truth .
With these expectations , I have ventured to reduce the method into its present form ; and submit it , with the utmost deference , to this learned Society .
In the'spherical triangle , whose sides are the complements of the latitude , declination , and altitude , let z represent the angle at the pole , and t its tangent ; Z the azimuth , and T its tangent ; L the latitude , and , , its cosine , radius being unity ; then , if the altitude and declination remain constant , we shall have L= TA 1 , and , consequently , L will vary as T - , when the increment of A , compared with x itself , is inconsiderable .
H-lence , if the abscisse of the curve ABCD ( fig. i. ) be always proFi .
G. portional to z , and its ordinate to T , the area GB intercepted betwixt any two of these ordinates may represent the increment of the latitude corresponding to the increment of the time EG .
Let abcdd ( fig. 2 .
) be .
another curve , whose abscisse ae is always equal to AE in the preceding figure , but whose ordiA 3 nate eb isl proportional to t , the tangent of the hour-angle ; then will the area gb vary FiG..as GB , at small distances from the meridian , and , of course , may represent the increment of the latitude .
Now , to prove this , we have only to show that T and t , when both are small , bear to each other a given ratio .
age Let S and Y be the sine and cosine of the azimuth ; s and T the sine and cosine of the angle at the pole ; then will Z. IFL ?
, and t-Z .
' , Z = ; - , and z_ But , since the '1 ' ~t 0* complements of the declination and altitude remain constant , whilst the latitude is made to vary , S will be to s as S to s ; TS+ TZ ~~~~~~I+ and , therefore , , we shall have:t::---x l w : X++ T1 : i+tl the square of the secant of the azimuth : the square of the secant of the hour-angle , which may be considered as a ratio of equality , when the angles are very small .
Tile fluxions , therefore , of the tangents are as the tangents themselves ; and , consequently , they must always preserve the same ratio towards each other .
Let us now suppose that an altitude of the sun is taken at the distance ae from the merit dian , but that , in consequence of an error in the assumed latitude , the calculated time is ag ; and that , with a lat. differing from the former by one minute , we compute-again , and the -time is found equal to a ; then will the area g , be to gb as on-e ranute to the whole error in latitud 6e .
L et another al titude be taken at the distance ae from -noon , and , let the times computed with the , two differeant latitiudes ' that were .
employed before be agi and af ; then , in'this case likewise ; the area.gc will be to the area gb , as one minute to the errorinlatitude .
' Now the latter curve is the ' " fgra .
tangenliur,7 ' whose quadrature is given by COTES , in his Harnmonia M3ensurarum , and the expression for which is extremely simple .
" For the fluxion of the area is , and the area itself log . ?
1-o og. .
Secant of the angle at the pole .
The difference of the log. .
secants , or log .
cosines , wili of course be equal to the area intercepted betwixt the tangents which correspond to them .
Hence a table might easily be constructed with a double argument , the distance from noon , and the variation in time arising from the different suppositions of latitude , -which might immediately exhibit the logarithm .
of the area corresponding to any particular base eg supposed to be given .
A second .
table might have for its argument the difference of the logarithms of the area gb and the area gi , ( which is likewise conceived to be known ) and discover at once , in degrees , minutes , and seconds , the correction .
to be made in the assumed latitude .
This correction , as it appears from a comparison of the signs of L and .
; in the equation L=XT m , must be added or subtracted , according as the distance from noon obtained by computation is too great or too .
little , when the azimuth is less than 9o degrees ; but the contrari , when the azimuthi exceeds a right " angle Tables.of the above description shall beconstructed , if thi method8 beMret ceived with approbation ; and , in the mean time , it is : proposed to subjoin a short specimen which is already -completed .
I have presumed that we are able to , ,determine , eg the error in time arising from an error in the ss2umeI latitude , at either of the observations and hence it becomes necessary , before we can avail ourselves of the principles which have been laid down , to point out the manner in which this ray be accomplished .
T ; Lhe clock gives us the whole interval betwixt the observations ( supposed L to be ; lmade on .
different sides of the meridian ) equal to ae + ae , and by computation ; we -obtain ag + -g , and thence we deduce eg + eg the whole error in time .
Now the area.gb is equal to gb ; ^aand therefore , if we make atrough division of the whole error , without any regard to accuracy , in the inverse ratio of the hour-angles at the two observations , and , .
entering the first table with thesetimes , mark the .
area corresponding to each at their respective distances ; from nloon , and increase the one and diminish the other equally , till we 'get the areas of the same magnitude , this , we may conclude , is the prop per value of each .
&gt; If the table were constructed to every second of time , we might ascertain these ogarithmic areas merely from inspection ; but , as it will be advisable to confine it within narrower limits , we shall sometimes find it necessary , as in other tables , to deduce their ultimate value by the rule of three .
When we have increased one-portion of time and diminished the other , till the difference of their corresponding areas becomes a mznimum , we must divide this difference betwixt them , in the proportion Of their respective increments in the next interval of time , and sub .
tract or add the part assigned to each , according as it is greater or less than the other .
The table , however , m-ight easily be carried to such an extent , that exactness in this division could never be required ; but , on the contrary , it would be quite sufficient , when the hour-angles were nearly equal , to add the areas together , and take half the sum for the value of each .
From these principles may be deduced the following practical rule for determining the latitudeof a place .
When the sun comes within fifteen degrees of the meridian , in the morning , let his altitude be taken , and the time of the observation be accurately marked ; and let another altitude be taken after he has passed the meridian , whilst his distance from it is less than fifteen de .
grees ; and let the -time of this observation likewise be noted .
Then , with the supposed latitude of the place , compute the times corresponding to each of the altitudes in terms of the log .
cosi of the hour-angle , and take:the difference of the intervals , shown by the clock , and determined by calculation , and divid , it betwixt the observations in the manner explained above .
Compute the log , cosine of the hour-aIngle a second= time , with the greatest altitude and the latitude increased or diminished by a minute , according as it appears , from a comparison of the intervals , to have been too little or too great ; and take the differencel betwixt this log .
cosine and that which resulted from the first operations , when the same altitude was employed .
Having thus obtained the two -areas gb and Lg(c , we must subtract their logarithms from -each other , and with their difference : entering the second table we shall findtheA degres , minutes , and seconds , by which the assumed latitude is to be increased or diminished .
It will -be needless , perhaps , to suggest that , in the higher latitudes ; we may extend the limits above specified a few degrees farther from the meridian , without offering any mantrial violence to the theory , as it has hitherto been explained ; x-and that , on the other hand , when the declination and latitude are nearly equal and of the same denomination , it whill be expedientto confine our observations within a : much shorter-distance from noon .
But it will afterwards be demonstrated that , 'whatever be the magnitude of the hour-angles , or however nearly:the latitude and declination -may approach towards each other , we can always secure , with very lifttle additional trouble , anexact conclusion .
We may remark that the latitude , determined in this mmanner , will be nearly equivalent , in point of accuracy , to the mean result of two meridian altitudes .
For we know that the increment of latitude : incrementof altitude:radius':cosine of azimuth ; and , since the cosine of asmall a-ngle differs so little from the radius , this may be c'o-nsidered , 's within the limits which I have prescribed , as a ; ratio of equality .
If , therefore one altitude of the sun were taken , and we.vould ascertain the error in time arising from an error in the .
assumed latitude without the aid of a second observation , the latitude would be discovered with nearly the sa'ie precision as if it had been deduced from the meridian altitude .
But , by means ot a second obseravationr made on a different side of noon , we obtain a second error fin ; time of the samekind ; and : this being added to the former , and their sum divi ded in , - : a just Iroportion hetwixt the two observations , the same effect wilblhe produced , with respect to the accuracy of the result , as if ' two latitudes : had : beendeduced from meridians adltitu-des , an1-d a , mean be-twixt them had been taken .
I might perhaps be allowed to say more ; forI am satisfied , from experience , that I can take an altitude of the .
sn with greater'exactness , When he is in any other situation , than when he is upon the meridian .
If we could ascertain , within a few seconds , or even within a minute , thetimne when he attains his greatest altitude , there : would thenv be no reason .
why an obServatioon should not be -made with the same degreeiof cervtainty in this , ,-as in/ other cases ; -but we are generally obliged tokeep our eye steadfastly fixed , for several minutes , upon the two images ; .
, and it is well known that , in such circumstances , the best eyes are apt to be deceived .
Besidesj it is impossible to 'preserve the contact of the limbs by perpetually moving the index , whilst the sun continues to ascend so very slowly .
We are compelled to wait till they are evidently separated , and then , by one turn of the screw , to bring them into contact again , which must necessarily be a source of some inac.:curacy .
It is for the first of these reasons that , in taking an altitude of the sun , when he is near the meridian , I have found it advisable , .
hot , inhthe usual manner , to bring the images almost to touch each : other , and then to wait till they actually do so , but to bring them at once into contact , with such a degree of velocity as would make them sensibly overlap , or separate , whilst the clock beats a second .
But I consider it as one-of the principal advantages of this method , that we can availourselves of any number of altitudes , and , of course , approximate as near as we please to a true conclusion with so little additional labour .
If there be an equal number of observations made on each side of the meridian , we , must combine them together by pairs , according to the preceding instructions , and thus determine the different logarithmicvvalues of gb .
Having then added them all together , and taken a mean betwixt them , we have only to come put a single incremental area gc with any of the altitudes and the lat. varied one minute , and , subtracting its log , from the mean log.-value ' of gb , we shall obtain a very accurate correction of the assumed latitude .
But , if there be more observations on one side of the meridian than on the other , when all the pairs have been united , and theareas resultingfrom them found , we may combine the supernumerary observations on either side with any of those which are made on the opposite side .
The properest , however , for this purpose , is the observa tion which is made at the least distance from the meridian .
I should hope , moreover , the practical astronomer will think it a circumstance of some moment , that the principal part of the work consists in finding the time , an operation which he is obliged so frequently to perform .
Any of the three methods which are usually adopted upon this occasion might easily be applied to the tables which have been described ; but I " iII venture to recommend a different rule , which I conceive to be better adapted to our purpose than any of the others , and to which the'directions before given had a particular reference .
Let a be the sine of the altitude ; y the cosine of the hourangle ; d the sine , &amp; the cosine , and r the -tangent of declination ; I the sine , X the cosine , and s the tangent of the latitude .
a-dl a dl _ ars a Then= = .
ars -r 5 S. , when radius is unity , but i rs when radius is m x. into the square of the tangent of the arc whose se : 3 4cant is d o. Hence we deduce the following rule for deters , mining the log , cosine of the angle at the pole .
From the log .
sine of the altitude increased by three times the log , radius , ~sulbtract the sum of the log .
sines of the latit .
and dedination ; take half of the remainder and , coMsidering it as the log , secant of an arc , find the log , .
tangent corresponding ; multil this by 2 , and -add the log .
, tangents of the latit .
and decline .
and reject thrice the log .
radius ; the sum will be the log .
cosine of the angle required .
But , when the decline .
and latit .
are of a different denomination , it is evident that our expression becomes m+ ad/ .
l3 , which is equal to 4-X , into the square of the secant of the arc whose tangent 'is fr &lt; .
In this case , therefore , having found the log , value of a and divided it by 2 , we must consider the quotient as the log .
tangent of an arc , whose log .
secant being taken , we are to proceed as in the former case .
The advantages of this rule are obvious .
We obtain the angle in terms of the log .
cosine , ; and , consequently , when we have calculated the second time with the new latit .
we , have only to subtract one result from the other , and we immediately determine the area corresponding to the difference of the times .
Besides in the second computation , fewer of the elements are changed by this rule , than by any of those which are usually employed ; and this is a consideration of much imp portance .
But , if we are disposed to adopt the following method of ascertaining the .
in cremental area gc , this advantage will be found still greater .
Let us resume the expression.a-dl I dx ia ; ( Ad be ~ and we shal have tie ( be+ dI ing the succeeding value of x ) =x dAxz-a Ix d1A dxx a-dl I ( taking x positive instead of negative , as it ought to be when i is positive ) s -s + and , consequetly , 21 2Z S when radius is unity , btit-= a2 i. v when radius is in .
Sy NI Now may be considered as the increment of the hyperbolic log .
of y , and therefore , with its proper modulus , may represent the area which is the object of our investigation .
We may suppose the other side , of the equation to be the square of the cosine of the : arc whosegsine is 9'i x. into the increment of the hyperbolic log .
of AX , divided by the square of the radius ; and if , instead of taking this log .
with the hyperbolic we take it with BRIGGS'S modulus , we must then consider Y as the increment of the log .
of y , according to the same system .
But.1 being equal to -s(when L is only one minute ) , it will vary as s ; and , therefore , if its value be determined according to BRIGGs 's system , When s is equal to radius , and be denominated v , its value in any other case will be expressed by v2$4Hence , to obtain the log , of the area gC , the quantity with which we are immediately concerned , we must find the log .
value of.- , and divide it by .
; we must then take out the log .
, cosine of the arc whose log .
sinie is equal to the quotient ; and , having multiplied it by owe must add the product to the constant log , of v ( 310ol5 ) and the , log tangent of the supM 2 posed latitude , rejecting thrice the log .
radius .
But , if rmZ be greater than radius , which must necessarily be the case when the azimuth is greater than a right angle , we must then consider Z ml ' as the square of the tangent of the arc Sr whose secant is 5 , observing in other respects the directions before given .
The quantities r and s are both employed in the first computation , from the result of which we also obtain y ; and , consequently , this operation will not be attended with much trouble .
The above instructions , it is manifest , are given upon the supposition of r and S having the same sign but , if the decline . .
and latit .
should not be of a similar .
denomination , then will our expression becomem1+ m. and we must consider ?
fl2 + rS 1as the square of the secant whose corresponding Sy tangent is rm .
With this exception , the process will be the same as when the .
tangents r and s are both affirmative .
The preceding formula naturally suggests to us another method of finding the log .
area gc ; and , as someperhaps may think this more eligible than either of the former , I shall take the liberty of explaining it .
When the latit .
-is given , the area G C , it is obvious , must invariably preserve the same magnitude at all distances from the meridian ; and , consequently , the area gc , which is proportional to it , must likewise remain constant .
If , therefore , we can ascertain this area .
when the hour-angle is supposed to vanish , we may employ it when the sun is at any distance .
from noon. .
Let us now conceive the decline .
to be equal to nothing ; then will our expression for the area gc become A and , consequently , ( since the tangents of the azimuth and hour-angle vanish in the ratio of their sines , or of the sines of the opposite sides in the triangle alluded to before , ) we shall have the area GC = t , when the sun is upon the meridian .
But this area is always the , same when the latitude is given , whatever be the sun 's declination , and therefore may always be represented by AV Xvi and the area g " will be generally expressed by -x co.s .
of merid .
galtit .
-i .
x cos. of med.ltit when the hour-angle does not exceed the limits cos. of decline .
which have been recommended .
Hence , if we add together the constant log .
3,1o15 , the log .
radius , and the log .
cosine of the merid .
altitude , and ' subtract from their sum the log .
cosines of the latit .
and decline .
we shall obtain the log .
value of gc .
It will be necessary , perhaps , to meet anl objection which some may be inclined to urge against the method of deducing the hour-angle in terms of..the cosine , when this angle is very small .
But it should be recollected , that with the angle itself we have no immediate concern , the accuracy of our conclusion depending entirely upon the accuracy with which the area corresponding to any particular incrrement of time can be determined .
Now this area , whatever be the sun 's distance fMrm the meridian , will be nearly proportional to the increment of the latit .
and , consequently , its magnitude is totally unconnected with that of the hour-angle .
A given error in the quantity which expresses thisarea will equally affect our conclusion , whether the angle be2 , or whether it be 2o degrees .
But let us inquire what effect will actually be produced , by admitting an error of half an unit in each of the log.-cosines whose difference is equal to the area gc ; and , of course , in some instances , an error of an unit in the area itself , upon any partiecular supposition of latit .
and decline .
We have only to-ascertain the ratio which this area bears to unity ; for the same ratio will the correction of the latit .
bear to the error in our re-sult .
If the latit .
for instance , be 50o and the ; decline lo , on the same side of the equator .
with the latit .
then , radius .
being L unity , Z ( the increment of the hour-angle ) will equal ~rp ( g being the sine of the hour-angle , and Cos. 5 0oX L9 the cos. of the altit .
) sx i= t minutes nearly , when z is t " ; and we have seen that , at any other distance from the meridian , the incremental area will -be of the same magnitude .
Hence , subtracting the log .
cosine of SO from that of 5c 11 ' o " , we get the difference equal to 1238 ; and , consequently , the error in our approximation will be to the error in the assumed latit .
as I1 9238 , when the log .
cosines are carried to seven places of decimals .
But when the zenith distance of the sun , at his greatest altit .
is very small , and there is moreover a considerable uncertainty with respect to the latit .
this error will probably become of more importance , and we may find it necessary to guard against it .
Now it is manifest that , by diminishing the multiple which the area gb is of the area gc , exactly in the same proportion we shall diminish this error ; and I shall afterwards explain in what manner it may be accomplished .
The same expedient is also calculated to prevent another species of inac-o curacy , to which , under similar circumstances , our process is liable .
If we -are desirous of knowing how much our conclusionis affected by substituting the tangent of thehour-angle for the tangent of the azimuth , it may be done with the greatest facility .
We have seen , that the area GC ( fig. 1 .
) is to GB as one minute to the error in the assumed latitude .
But our method supposesthat gc ( fig. 2 .
) is to gb as one minute to the correction required ; and we must therefore estimate the difference betwixt these ratios , in order to ascertain how far the approximlation is inaccurate .
Let EG-m x FG ; and then , since GC may be considered as -FG , and GB 2 _D + GD +EB _GD + EB x EG _xmx FG , the former ratio will 2 GD ?
+ED equal Gxm= ( according to the notation , employed above ) m.:-T --=m_--m.m 2 nearly = the minutes contained in the computed error of latitud-e .
Their difference r-i 'T t I. T. + will equalm.- .
T-T=mm .
2* Z= .
ttI T2_ t2 z_ mrm i. T ?
2t , LI --being 2tz ATt T substituted for its equal)=the minutes contained !
in the error of our approximation .
It will be more convenient , however , to express this difference independently of the azimuth .
Now , preserving the notation before adopted , we Sg v7~ tyhrfr have T= -_d-t d7 therefore , TT VI _dT_ av~ ' Xy -d dA x IXy_ dt being rejected , as incomparably less than the sum of the other terms , ) X. A9d !
66= t*d nsequently , we I &lt; d s-yr m -I Tz-t in-i '29 +2 r havew .
MT Lm2-.-Z ?
L lni .
2 ATt L 100000 ' sl-r nutes .
Hence it appears that , in the latitude of Cambridge , when the sun 's declination is ?
north , and his distance from in -I -i the meridian ?
, this error will be equal to m.:2 minutes of a degree .
Let the assumed diff from^ the true latitude ten minutes , z. e. let mi a , then Qwill the error amount to half a second ; and , if we suppose ma 3o , the error will not exceed four seconds .
If , instead of varying the assumed latitude one minute , we vary it n minutes , in calculating the area gc , the above ex11 ~ ~~~~~~ 2 , -m I Zgr .
n 2 .
, sr s pression will be transformed into m -=- , _zL , -,2L 2r 1 00000 Sr the increment of the latitude being , in this case , n L. The error will consequently vary as m m. n , when the declinaation , latitude , and time of observation are given ; and .
, if we suppose the real to differ from the assumed latitude p minutes , p -~~~~.-I p npn this last expression will become P , n-p .
7 ' which varies as it ?
nn when p remainsconstant ; and , of course , n the error may be diminished in any proportion by diminishing this quantity PL-n , Now , the increment of the latitude is always xT z ; and , therefore , since z is determined by the process explained above , we have only to ascertain the value of T , in order to approximate nearly to p. But the sine of the

