1. If `lim_(x->c) f(x) g(x) ` exists, then we can have the following cases:
`quadquadquadquad (a)` Both `lim_(x->c)f(x)` and `lim_(x->c) g(x)` exist. Obviously, then `lim_(x->c) f(x) g(x)` exists.
`quadquadquadquad (b)` `lim_(x->c) f(x)` exists and `lim_(x->c)g(x)` does not exist.
Consider `f(x) = x; g(x) = 1/x`, now `lim_(x->0)f(x) *g(x)` exists `=1`. Also `lim_(x->0)f(x) =0` exists but
`lim_(x->0)g(x)` does not exist.
`quadquadquadquad (c)` Both `lim_(x->c)f(x) ` and `lim_(x->c)g(x)` do not exist.
Let `f` be defined as `f(x) = {tt( (1, text(if) x le 0) ,(2, text(if) x >0) )` . Let `g(x) ={tt( (2, text(if) x le 0) ,(1, text(if) x>0) )`
Then `f(x) g(x) = 2`, and so `lim_(x->0) f(x) xx g(x)` exists, while `lim_(x->0)f(x)` and `lim_(x->0)g(x)` do not exist.
2. If `lim_(x->c) [ f(x) +g(x)]` exists then we can have the following cases:
`quadquadquadquad (a)` If `lim_(x->c) f(x)` exists, then `lim_(x->c)g(x)` must exist.
`text( Proof: )` This is true as `g = (f+ g) - f`.
Therefore, by the limit theorem, `lim_(x->c) g(x) =lim_(x->c) (f(x) +g(x)) - lim_(x->0)f(x)` which exists.
`quadquadquadquad (b)` Both `lim_(x->c)f(x) ` and `lim_(x->c)g(x)` do not exist.
Consider `lim_(x->1) [x]` and `lim_(x->1)g[x]`, where `[.]` and `{.}` represent greatest integer and fractional part function, respectively. Here both the limits do not exist but `lim_(x->1)[x]+{x} = lim_(x->1) x=1` exists.
`text(For Example:)`
(a) `f(x) =1/sin x ` and `g(x) =1/tan x` at `x=0` `Lim_(x->0) f(x) = DNE`, `Lim_(x->0)g(x) =DNE`
But `Lim_(x->0 ) (1/sinx -1/tanx) =0 =Lim_(x->0) (1-cos x)/sin x`
`quadquadquadquadquadquadquadquad= Lim_(x->0) (2 sin^2 (x/2))/(2 sin (x/2 )cos (x/2)) `
`quadquadquadquadquadquadquadquad= Lim_(x->0) tan (x/2) =0` (exist).
(b) `f(x) = sgn x` and `g (x) = [x]` then `Lim_(x->0) (sgn x + [x])` does not exist
as `Lim_(x->0^+) sgn x + [x] =1` , `Lim_(x->0)sgn x +[x] =-2`
while `Lim_(x->0) f(x) =DNE`, `Lim_(x->0) g(x) =DNE`
(c) `f(x) = [x] and g(x) = {x}; F(x) = [x] {x}`
`Lim_(x->0) [x] *{x} ` does not exist but `Lim_(x->1)[x]{x}` exist and is equal to zero.
(d) If `f(x) = e^( [ x ])` ; `g(x) = e^({x}) ` then `Lim_(x->0) e^ ([x]) * e^( {x}) = Lim_(x->0) e^x` which exist.
1. If `lim_(x->c) f(x) g(x) ` exists, then we can have the following cases:
`quadquadquadquad (a)` Both `lim_(x->c)f(x)` and `lim_(x->c) g(x)` exist. Obviously, then `lim_(x->c) f(x) g(x)` exists.
`quadquadquadquad (b)` `lim_(x->c) f(x)` exists and `lim_(x->c)g(x)` does not exist.
Consider `f(x) = x; g(x) = 1/x`, now `lim_(x->0)f(x) *g(x)` exists `=1`. Also `lim_(x->0)f(x) =0` exists but
`lim_(x->0)g(x)` does not exist.
`quadquadquadquad (c)` Both `lim_(x->c)f(x) ` and `lim_(x->c)g(x)` do not exist.
Let `f` be defined as `f(x) = {tt( (1, text(if) x le 0) ,(2, text(if) x >0) )` . Let `g(x) ={tt( (2, text(if) x le 0) ,(1, text(if) x>0) )`
Then `f(x) g(x) = 2`, and so `lim_(x->0) f(x) xx g(x)` exists, while `lim_(x->0)f(x)` and `lim_(x->0)g(x)` do not exist.
2. If `lim_(x->c) [ f(x) +g(x)]` exists then we can have the following cases:
`quadquadquadquad (a)` If `lim_(x->c) f(x)` exists, then `lim_(x->c)g(x)` must exist.
`text( Proof: )` This is true as `g = (f+ g) - f`.
Therefore, by the limit theorem, `lim_(x->c) g(x) =lim_(x->c) (f(x) +g(x)) - lim_(x->0)f(x)` which exists.
`quadquadquadquad (b)` Both `lim_(x->c)f(x) ` and `lim_(x->c)g(x)` do not exist.
Consider `lim_(x->1) [x]` and `lim_(x->1)g[x]`, where `[.]` and `{.}` represent greatest integer and fractional part function, respectively. Here both the limits do not exist but `lim_(x->1)[x]+{x} = lim_(x->1) x=1` exists.
`text(For Example:)`
(a) `f(x) =1/sin x ` and `g(x) =1/tan x` at `x=0` `Lim_(x->0) f(x) = DNE`, `Lim_(x->0)g(x) =DNE`
But `Lim_(x->0 ) (1/sinx -1/tanx) =0 =Lim_(x->0) (1-cos x)/sin x`
`quadquadquadquadquadquadquadquad= Lim_(x->0) (2 sin^2 (x/2))/(2 sin (x/2 )cos (x/2)) `
`quadquadquadquadquadquadquadquad= Lim_(x->0) tan (x/2) =0` (exist).
(b) `f(x) = sgn x` and `g (x) = [x]` then `Lim_(x->0) (sgn x + [x])` does not exist
as `Lim_(x->0^+) sgn x + [x] =1` , `Lim_(x->0)sgn x +[x] =-2`
while `Lim_(x->0) f(x) =DNE`, `Lim_(x->0) g(x) =DNE`
(c) `f(x) = [x] and g(x) = {x}; F(x) = [x] {x}`
`Lim_(x->0) [x] *{x} ` does not exist but `Lim_(x->1)[x]{x}` exist and is equal to zero.
(d) If `f(x) = e^( [ x ])` ; `g(x) = e^({x}) ` then `Lim_(x->0) e^ ([x]) * e^( {x}) = Lim_(x->0) e^x` which exist.