less than q3/ 27 but greater than q3/ 54 , and e be put = r/ 2 , and zz be put = q3/ 27 rr/ 4 , the root of the equation x3 qx =r will be equal to ~3 ex the infinite series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. ART .
4 .
The numeral coefficients 2/ 9 , 20/ 243 , 308/ 6561 , &amp;c. of zz/ ee , z4/ e4 , z6/ e6 , &amp;c. in this series are exactly double of 1/ 9 , 10/ 243 , 154/ 6561 , &amp;c. which are the numeral coefficients of the same powers of the fraction z/ e in the series 1+ z/ 3e zz/ 9ee + 5z3/ 81e3 10z4/ 243e4 + 22z5/ 729e5 154z6/ 6561e6 + 2618z7/ 137,781e7 &amp;c. which is equal to the cube-root of the binomial quantity 1+ z/ e ; or , if the numeral coefficients of the said latter series be denoted by the capital letters A , B , C , D , E , F , G , H , &amp;c. respectively , so that A shall be = 1 , B= 1/ 3 , C= 1/ 9 , D= 5/ 81 , E= 10/ 243 , F= 22/ 729 , G= 154/ 6561 , and H= 2618/ 137,781 , and so on , the said numeral coefficients 2/ 9 , 20/ 243 , 308/ 6561 , &amp;c. will be equal to 2C , 2E , 2G , &amp; C. and the series mentioned in the last Article will be 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. and consequently the root of the equation x3 qx = r , in the second case of it , in which rr/ 4 is less than q3/ 27 , will be equal to the expression ~3 ex the series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. ART .
5 .
Now the series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. is equal to the sum of the two following serieses , to wit , 2A 2Czz/ ee 2Ez4/ e4 2Gz6/ e6 &amp;c. ( in which all the terms following the first term are marked with the sign - , or are subtracted from the first term ) , and 4Czz/ ee + 4Gz6/ e6 + &amp; c ; and the series 2A 2Czz/ ee 2Ez4/ e4 2Gz6/ e6 &amp;c. is equal to the sum of the two serieses A+ Bz/ e Czz/ ee + Dz3/ e3 Ez4/ e4 + ~ Gz6/ e6 + &amp;c. and A Bz/ e Czz/ ee Dz3/ e3 Fz5/ e5 Gz6/ e6 &amp;c. which are respectively equal to the cube-roots of the binomial quantities 1+ z/ e and 1 z/ e. Therefore the series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. is = ~3 1+ z/ e + ~3 1 z/ e + the infinite series 4Czz/ ee + 4Gz6/ e6 + &amp;c. Consequently the expression ~3 ex the series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. is = ~3e x ~3 1+ z/ e + ~3e x ~3 1 z/ e + ~3e x the infinite series 4Czz/ ee +~+ &amp;c. = ~3 e+z+ ~3 ez+ ~3 ex the infinite series 4Czz/ ee + 4Gz6/ e6 + &amp;c. = ~3 e+z+ ~3 ez+ 4 ~3 ex the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz18/ e18 + &amp;c. Therefore the root of the equation x3 qx = r , in the second case of it , in which rr/ 4 is less than q3/ 27 , is equal to ~3 e+z+ ~3 ez+4 ~3 ex the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz14/ e14 + &amp; c ; of which expression the first part , to wit , ~3 e+z+ ~3 ez is algebraick , and the latter part , to wit , ~3 ex the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz18/ e18 + &amp;c. is transcendental .
Q. E. I. ( a ) Of the convergency of the Series obtained in the preceding Article .
ART .
6 .
This series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz18/ e18 + &amp;c. evidently converges faster than the series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 &amp;c. or 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp; c : and consequently the expression ~3 e+z+ ~3 ez+ 4~3 ex the series Czz/ ee + Gz6/ e6 + Lz10/ e1~ + Pz14/ ~ + Tz18/ e18 + &amp;c. seems rather fitter ( a ) N. B. I have been informed that both this mixed expression of the root of the equation x3 -qx =r in the second case of it , and the merely transcendental expression of it published in the former paper , and from which this expression is derived , were invented by Monsieur NICOLE , and published in the memoirs of the French Academy of Sciences so long ago as the year 1738 ; and the latter of them , to wit , the transcendental expression ~3 ex the series 2+ 2ss/ 9ee 20s4/ 243e4 + 308s6/ 6561e6 &amp;c. I had myself seen many years ago in Monsieur CLAIRAUT'S algebra , in the place cited in the 50th Article of my former paper , to wit , in pages 286 , 287 , 288 .
But it was obtained by the intervention of negative quantities , and the roots of negative quantities , which gave it , in my opinion , an air of great obscurity .
And therefore I thought an investigation of the same series , by a method that keeps clear of those difficulties , might not be unacceptable to the lovers of these sciences , nor unworthy of a place in the Transactions of this learned body .
to exhibit the value of x in the equation x3-qx = r , in the second case of it ( in which rr/ 4 is less than q3/ 27 ) , to a considerable degree of exactness than the former expression ~ 3e X the series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. A Computation of the four first Terms of the Series obtained in Art .
5 .
ARE T. 7 .
The first fifteen terms of the infinite series which is equal to the cube-root of 1+ z/ e are as follows ; to wit , 1+ z/ 3e zz/ 9ee + 5z3/ 81e3 10z4/ 243e4 + 22z5/ 729e5 154z6/ 6561e6 + 2618z7/ 137,781e7 935z8/ 59,049e8 + 21505z9/ 1,594,323e9 55913z10/ 4,782,969e10 + 147,407z11/ 14,348,907e11 1,179,256z12/ 129,140,163e12 + 3,174,920z13/ 387,420,489e13 60,323,480z14/ 8,135,830,269e14 + &amp; c ; or , in decimal fractions , 1+ .333,333,333 , &amp;c. X z/ e -.111,111,111 , &amp;c. X zz/ ee + .061,728,395 , &amp;c. X z3/ e3 -.041,152,263 , &amp;c. X z4/ e4 + .030,178,326 , &amp;c. X z5/ e5 -.023,472,031 , &amp;c. X z6/ e6 + .019,001,167 , &amp;c. X z7/ e7 -.015,834,305 , &amp;c. X z8/ e8 + .013,488,482 , &amp;c. X z9/ e9 -.011,690,017 , &amp;c. X z10/ e10 + .010,273,045 , &amp;c. X z11/ e~ -.009,131,595 , &amp;c. X z12/ e12 + .008,195,021 , &amp;c. X z13/ e~ -.007,414,542 , &amp;c. X z14/ e14 .
Therefore the four first terms of the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz18/ c18 + &amp;c. are zz/ 9ee + 154z6/ 6561e6 + 55913z10/ 4,782,969e10 + 60,323,480z14/ 8,135,830,269e14 , or , in decimal fractions,.111,111,111 , &amp;c. X zz/ ee + .23,472,031 , &amp;c. X z6/ e6 +.011,690,017 , &amp;c. ~ z10/ e~ + .007,414,542 , &amp;c. X z14/ e14 .
Therefore the root of the cubick equation x3 qx = r , in the second case of it ( in which rr/ 4 is less than q3/ 27 ) , is equal to ~3e+z+~3ez+ 4~3e X the series zz/ 9ee + 154z6/ 6561e6 + 55,913z10/ 4,782,969e10 + 60,323,480z14/ 8,135,830,269e14 + &amp;c. ad infinitum , or ~3e+z+~3ez+4~3eX the series .111,111,111 , &amp;c. X zz/ ee +.023,472,031 , &amp;c. X z6/ e6 + .011,690,017 , &amp;c. X z10/ e~ +.007,414,542 , &amp;c. X z14/ e~ + &amp;c. ad infinitum .
Of the best Manner of Proceeding to the Computation of more Terms of the said Series , if required .
ARE T. 8 .
If more than four terms of this last series are required , it will be necessary to compute the series 1+ z/ 3e zz/ 9ee + 5z3/ 81e3 10z4/ 243e4 + 22z5/ 729e5 154z6/ 6561e6 + &amp;c. ( which is =~31+ z/ e ) , to more than fifteen terms ; in order to which it will be convenient to express the terms of that series in the following manner , to wit , 1+ 1/ 3 X Az/ e 2/ 6 X Bzz/ ee + 5/ 9 X Cz3/ e3 8/ 12 X Dz4/ e4 + 11/ 15 X Ez5/ e5 14/ 18 X Fz6/ e6 + 17/ 21 X Gz7/ e7 20/ 24 X Hz8/ e8 + 23/ 27 X Iz9/ e9 26/ 30 X Kz10/ e10 + 29/ 33 X Lz11/ e11 32/ 36 X Mz12/ e12 + 35/ 39 X Nz13/ e13 38/ 42 X Oz14/ e14 + 41/ 45 X Pz15/ e15 44/ 48 X Qz16/ e16 + 47/ 51 X Rz17/ e17 50/ 54 X Sz17/ e~ + 53/ 57 X Tz~/ e~ 56/ 60 X Vz~/ e20 + 59/ 63 X Wz21/ e21 62/ 66 X Xz22/ e22 + &amp;c. or A+ Bz/ e Czz/ ee + Dz3/ e3 Ez4/ e4 + Fz5/ e5 Gz6/ e6 + Hz7/ e8 + Iz8/ e8 + K~i/ e9 Lz10/ e10 + Mz11/ e~ N12/ e12 + ~z13/ e13 Pz14/ e14 + 41/ 45 X Pz15/ e15 44/ 48 X Qz16/ e16 + 47/ 51 X Rz17/ e~ 50/ 54 X Sz18/ e18 + 53/ 57 X Tz19/ e19 56/ 60 X Vz20/ ~20 + 59/ 63 X Wz21/ e21 62/ 66 X Xz22/ e22 + &amp; c ; in which it is evident that the generating fractions of the coefficients of the several terms are derived from those that immediately precede them by the continual addition of the number 3 to both their numerators and denominators .
An example of the resolution of a cubick equation by means of the expression ~3e+z+~3ez+4~3eX the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e~ + Tz18/ e~ + ~c .
given in Art .
5 .
ARE T. 9 .
Let it be required to resolve the equation x3 x= 1/ 3 by means of the said expression .
Here q is = 1 , and r= 1/ 3 ; and consequently q3 is = 1 , and q3/ 27 = 1/ 27 , and r/ 2 = 1/ 6 , and rr/ 4 = 1/ 36 , which is less than 1/ 27 , or q3/ 27 .
Therefore this equation does not come under CARDAN'S rule , but may be resolved by the expression given in Art .
5 , provided that rr/ 4 , though less than q3/ 27 , is greater than half q3/ 27 , or than q3/ 54 ; which it is , because it is = 1/ 36 , whereas q3/ 54 is equal only to 1/ 54 , which is less than 1/ 36 .
Therefore the proposed equation x3 x= 1/ 3 may be resolved by means of the said expression .
ART .
10 .
Now , since in this case q is = 1 , and r is = 1/ 3 , we shall have r/ 2 , or e , =~ and q3/ 27 rr/ 4 ( = 1/ 27 1/ 36 = 4/ 108 3/ 108 = 1/ 108 ) = 1/ 36 x 3 ; that is , zz will be = 1/ 36 x 3 , and consequently z will be = 1/ 6 ~3 .
Therefore e+z will be = 1/ 6 + 1/ 6 ~3 ( = ~ 3/ 6 ~3+ 1/ 6 ~3= 3/ 6 x3+~ 3/ 6 x3=3+~ 3/ 6 x3=3+~ 3/ 18 =3+ 1.732,050,8/ 18 = 4.732,050,8/ 18 ) =.262,891,71 ; and ez will be = 1/ 6 1/ 6 ~3 ( = 3~ 3/ 18 =3 1.732,050,8/ 18 = 1.267,949,2/ 18 ) =.070,441,62 .
Therefore the cube-root of e+z is =~ 3.262,891,71 = .640,607,91 ; and the cube-root of ez is = ~ 3 .
070,441,62 =.412,993,40 ; and consequently ~3e+z+~3ez is =.640,607,91 +.412,993,40= 1 .
053,601,31 .
ART .
II .
It remains that we compute the infinite series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + Tz16/ e16 + &amp;c. and extract the cuberoot of e , and then multiply the said series into 4 times the said cube-root .
Now the cube-root of e is in this case =~3 1/ 6 = 1/ ~36 = 1/ 1.817,121 ; and consequently 4~3e is = 4/ 1.817,121 .
And , since zz is = 1/ 36 x 3 , and ee is = 1/ 36 , it follows that zz/ ee will be = 1/ 3 .
Therefore z4/ e4 will be = 1/ 9 , and z6/ e6 = 1/ 27 , and z10/ e10 ( =z6/ e6 x z4/ e4 = 1/ 27 x 1/ 9 ) = 1/ 243 , and z14/ e14 ( = z10/ e10 x z4/ e4 = 1/ 243 x 1/ 9)= 1/ 2187 .
Therefore the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + &amp;c. will in this case be = C/ 3 + G/ 27 + L/ 243 + P/ 2187 + &amp;c. =.111,111,111 , &amp;c. / 3 + .023 , 472 , 031 , &amp;c. / 27 +.011 , 690 , 017 , &amp;c. / 243 +.007 , 414 , 542 , &amp;c. / 2187 + &amp;c. = .
037 , 037 , 037 , &amp;c. +.000 , 869 , 334 , &amp;c. + 000 , 048 , 107 , &amp;c. + 000 , 003 , 390 , &amp;c. + &amp;c. + &amp;c. = .
037 , 957 , 868 , &amp;c. Therefore 4 ~3e x the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + &amp;c. is equal to 4/ 1.817,121 x. 037 , 957 , 868 , &amp;c. = .
151 , 831 , 472 , &amp;c. / 1.817 , 121 = .
083 , 556 , 06 .
Consequently the whole expression ~3e+z+~3ez+ 4~3ex the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + &amp;c. is = 1 .
053 , 601 , 31 + .
083 , 556 , 06 = 1 .
137 , 157 , 37 ; that is , the root of the proposed equation x3 x= 1/ 3 is = 1 .
137 , 157 , 37 .
Q. E. I. ART .
12 .
This value of the root of the equation x3 x= 1/ 3 is exact to six places of figures , the more accurate value of it being 1 .
137 , 158 , 164 .
We may therefore conclude that the expression here made use of to determine the value of x , to wit , ~3e+z+~3ez+ 4~ 3e x the infinite series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + &amp;c. is somewhat preferable , with respect to the practical resolution of these equations , to the other expression of its value given in the former paper in the Philosophical Transactions , to wit , ~3e x the infinite series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. For it appeared in Art .
42 of that paper , page 941 , that the value of the root of this same equation x3 x= 1/ 3 obtained by computing four terms of the series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. was 1 .
137 , 33 ; which is true only to four places of figures ; whereas by computing the same number of terms of the series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e 4+ &amp;c. we have just now obtained a value of the same root , to wit , the number 1 .
137 , 157 , 37 , which is exact to six places of figures .
This is agreeable to what was observed above in Art .
6 .
A Summary of the Conclusions obtained in this Paper and the former Paper to which it is an Appendix .
ART .
13 .
I will now conclude this paper by setting down , in as concise a manner as I can , the several conclusions that have been obtained in this and the abovementioned paper in the Philosophical Transactions for the year 1778 , Number XLII .
page 902 , &amp;c. concerning the root of the cubick equation x3 qx = r , that the whole may be seen together at one view .
ARE T. 14 .
If rr/ 4 is greater than q3/ 27 , and e be put = r/ 2 , and ss = rr/ 4 q3/ 27 , it is shown in Art .
5 , of the said former paper , that the root of the equation x3 qx =r will be =~3 r/ 2 +~ rr/ 4 q3/ 27 +q/ 3 ~3 ~ r/ 2 +~ rr/ 4 q3/ 27 , or ~3 ~e +s+ q/ 3 ~3~e+ s. ARE T. 15 .
And it is shown in Art .
9 .
of the said paper , that the root of the said equation will in that case be also equal to ~3~ r/ 2 ~ rr/ 4 q/ 27 q/ 3 ~3~ r/ 2 ~ rr/ 4 q3/ 27 , or ~3~es+ q/ 3 ~3~e s. ARE T. 16 .
And it is shown in Art .
11 , of the said paper , page 915 , that the root of the said equation will in that case be also equal to ~3~ r/ 2 +~ rr/ 4 q3/ 27 +~3~ r/ 2 ~ rr/ 4 q3/ 27 , or ~ 3e +s+ ~3 ~e s. ARE T. 17 .
And it is shown in Art .
23 , of the said paper , page 923 , that the root of the said equation will in that case be also equal to ~ 3e ~ the infinite series 2 2ss/ 9ee 20s4/ 243e4 308s6/ 6561e6 1870s8/ 59049e8 111,826s10/ 4,782,969e10 2,358,512s12/ 129,140,163e12 120,646,960s14/ 8,135,830,269e14 120,646,960s14/ 8,135,830,269e14 &amp;c. or ( if we put the capital letters A , B , C , D , E , F , G , H , &amp;c. for the several numeral coefficients 1 , 1/ 3 , 1/ 9 , 5/ 81 , 10/ 243 , 22/ 729 , 154/ 6561 , 26/ 137,18/ 781 , &amp;c. of the terms of the series 1+ s/ 3e ss/ 9ee + 5s3/ 81e3 10s4/ 243e4 + 22s5/ 729e5 154s6/ 6561e6 + 2618s7/ 137,781e7 &amp;c. which is equal to the cube-root of the binomial quantity 1+ s/ e ) , equal to ~ 3e ~ the infinite series 2A 2Css/ ee 2Es4/ e4 2Gs6/ e6 2Is8/ e8 2Ls10/ e10 2N12/ e12 2P14/ e14 &amp; c ; in which series all the terms after the first term.2A , or 2 , are marked with the sign - , or are to be subtracted from the said first term .
ART .
18 .
And , if rr/ 4 is less than q3/ 27 , but greater than its half , or q3/ 54 , and e be put , as before , = r/ 2 , and zz = q3/ 27 rr/ 4 , it is shown in Art .
31 , 32 , 33 , 34 , 35 , of the said paper , pages 927-936 , that the root of the equation x3 qx =r will be equal to ~ 3e ~ the infinite series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 1870z8/ 59049e8 + 111,826z10/ 4,782,969e10 2,358,512z12/ 129,140,163e12 + 120,646,960z14/ 8,135,830,269e14 &amp;c. or ~ 3e ~ the infinite series 2A + 2Czz/ ee 2Ez4/ e4 + 2Gz6/ e6 2Iz8/ e8 + 2Lz10/ e10 2Nz12/ e12 + 2Pz14/ e14 2Rz16/ e16 + 2Tz18/ e18 &amp; c ; in which series the capital letters A , C , E , G , I , L , N , P , R , T , &amp;c. denote the same numeral coefficients 1 , 1/ 9 , 10/ 243 , 154/ 6561 , &amp;c. as in the last Article , and all the terms that involve the odd powers of the fraction zz/ ee , to wit , zz/ ee itself , zz/ ee ~3 , or z6/ e6 , zz/ ee ~5 , or z10/ e10 , zz/ ee ~7 , or z14/ e14 , zz/ ee ~9 , or z18/ e18 , &amp;c. are marked with the sign + , or are to be added to the first term 2 A , or 2 , and all the other terms are marked with the sign - , or are to be subtracted from the former .
These are the conclusions obtained in the said former paper , which is printed in the Philosophical Transactions for the year 1778 , Number XLII .
pages 902 949 .
The conclusions obtained in this paper are as follows .
ART .
19 .
If rr/ 4 is less than q3/ 27 , but greater than its half , or q3/ 54 , and e be put , as before , = r/ 2 , and zz=q3/ 27 rr/ 4 , it is shown in the present paper , Art .
5 and 7 , that the root of the equation x3 qx =r will be equal to the mixed expression ~3 ~e+z+~3~ez+4 ~3 e~ the infinite series zz/ 9ee + 154z6/ 6561e6 + 55913z10/ 4,782,969e10 + 60,323,480z14/ 8,135,830,269e14 + &amp;c. or ~3e+z+ ~3~ez+4~3e~ the infinite series .
111,111,111 , &amp;c. ~ zz/ ee +.023,472,031 , &amp;c. ~ z6/ e6 +.011,690,017 , &amp;c. ~ z10/ e10 + .007,414,542 , &amp;c. ~ z14/ e14 + &amp;c. or ~3 ~e+z+~3~ez+ 4~ 3e ~ the infinite series Czz/ ee + Gz6/ e6 + Lz10/ e10 + Pz14/ e14 + &amp; c ; in which series the capital letters C , G , L , P , &amp;c. denote the same numeral coefficients 1/ 9 , 154/ 6561 , 55913/ 4,782,969 , 60,323,480/ 8,135,830,269 , &amp;c. as they denoted in the two last Articles , to wit , the coefficients of zz/ ee , z6/ e6 , z10/ e~ , z14/ e14 , z18/ e18 , &amp;c. or of the odd powers of zz/ ee in the series which is equal to ~3 ~1+ z/ e. And it is also shown in Art .
6 and 12 of this paper , that this expression , computed to a given number of terms , gives the value of x somewhat more exactly than the former expression , ~3e x the infinite series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. if computed only to the same number of terms .
ART .
20 .
As to the second branch of the second case of the equation x3 qx = r , or that in which rr/ 4 is less than half q3/ 27 , or than q3/ 54 , I do not know any method of extending CARDAN'S rule to it .
But I have been informed by my learned and ingenious friend Dr. CHARLES HUTTON , Professor of Mathematicks in the Royal Academy at Woolwich , that he has discovered such a method : and I hope he will soon communicate it to this learned V. An Appendix to the Paper in the Philosophical Transactions for the Year 1778 , Number XLII , pages 902 et seq. entitled , ` ` A Method of extending Cardan 's Rule for resolving one Case of the Cubick Equation x3 qx =r to the other Case of the same Equation , which it is not naturally fitted to solve , and which is therefore called the irreducible Case .
' ' By Francis Maseres , Esq. F. R. S. Cursitor Baron of the Exchequer .
Read Nov. 4 , 1779 .
ARTICLE I. IN the above-mentioned paper in the Philosophical Transactions the expression ~3ex the infinite series 2+ 2ss/ 9ee 20s4/ 243e4 + 308s6/ 6561e6 &amp;c. is shown to be equal to the root of the equation x3 qx = r , whenever rr/ 4 is less than q3/ 27 , but greater than one half of it , or than q3/ 54 .
This expression is wholly transcendental , or composed of an infinite number of terms , to wit , the terms of the series 2+ 2ss/ 9ee 20s4/ 243e4 + 308s6/ 6561e6 &amp;c. multiplied into the cuberoot of e. But I have since thought that it might be convenient on some occasions to divide this expression , if possible , into two others , whereof the one should be a mere algebraick expression , or consist of a finite number of terms , and the other should be transcendental , or involve in it an infinite series .
And I have accordingly discovered a method of doing this , which I will now proceed to describe .
ART .
2 .
In the above-mentioned paper in the Philosophical Transactions I denoted the excess of q3/ 27 above rr/ 4 in the second case of the equation x3 qx = r , as well as the excess of rr/ 4 above q3/ 27 in the first case of it , by the letters ss .
But I have since thought that it might have been better to denote the excess of q3/ 27 above rr/ 4 in the second case of that equation by the letters zz , in order the more clearly to distinguish it from the opposite difference rr/ 4 q3/ 27 in the first case of it , which was denoted by ss .
And I therefore in the course of the following pages shall use the letters zz instead of ss to denote the said excess of q3/ 27 above rr/ 4 in the second case of the said equation , or the difference q3/ 27 rr/ 4 .
ART .
3 .
Now , if zz be substituted instead of ss in the expression ~3e x the infinite series 2+ 2ss/ 9ee 20s4/ 243e4 + ~ &amp;c. that expression will thereby be converted into the following expression , to wit , ~3e x the infinite series 2+ 2zz/ 9ee 20z4/ 243e4 + 308z6/ 6561e6 &amp;c. Therefore , if rr/ 4 be

