Why Tables and Graphs Help
A sequence is an ordered list of terms indexed by integers, usually written \(a_1,a_2,a_3,\ldots\). To estimate its long-term behavior, we examine terms with increasingly large indices. A table makes their values easy to compare; a graph makes their overall pattern easier to see.
These tools help us form a reasonable estimate, but they have limits. A table shows only selected terms, and a graph shows only the terms that have been plotted. Neither one, by itself, proves what happens at every later index. The goal here is to read the evidence carefully and distinguish a likely pattern from a proved conclusion.
What a Sequence Graph Shows
A graph of a sequence plots points of the form \((n,a_n)\). The horizontal coordinate is an integer index, and the vertical coordinate is the corresponding term. Unlike the graph of a function defined for every real input, a sequence graph consists of separate points. Drawing lines between those points can make a pattern easier to see, but the connecting lines are only visual guides: they are not additional terms of the sequence.
For a graph, read the points from left to right, since that is the direction of increasing \(n\). Pay particular attention to points at large indices. Early terms may be far from the eventual pattern, and a vertical scale that is too wide or too narrow can disguise how much the terms change.
Estimating a Finite Limit
If the terms at larger indices cluster near a number \(L\), estimate that number as the sequence’s finite limit. In symbols, the estimate is written \(\lim_{n\to\infty}a_n=L\). Values need not equal \(L\): the important visual clue is that their distance from \(L\) seems to get smaller.
Worked Example: Estimating a Finite Limit from a Table
Consider \(a_n=3-\frac{1}{n+1}\). Compute terms at increasing indices and compare each one with 3.
The sampled terms rise toward 3 from below, and the displayed differences from 3 decrease from \(0.5\) to \(0.02\). A reasonable estimate is therefore \(\lim_{n\to\infty}a_n=3\). The calculation checks that each table entry matches the formula and that \(3-a_n=\frac{1}{n+1}\) at each listed index. The terms are not equal to 3 at any of these indices; the estimate concerns their trend, not an eventual equality.
There is a simple algebraic reason the estimate fits this example. The distance between a term and 3 is exactly \(\frac{1}{n+1}\). As \(n\) increases, the denominator increases, so this positive distance decreases toward 0. Thus the sequence’s terms get closer to 3. A table illustrates that relationship, while the expression for the distance explains why it continues beyond the sampled indices.
Estimating Growth Without Bound
A sequence that grows without bound does not approach a finite horizontal level. Its terms eventually exceed any height we choose, however large. A table may suggest this behavior when later terms keep rising, but the formula or a longer pattern of evidence matters: a few increases alone do not rule out a later turn or a leveling off.
Worked Example: Recognizing Unbounded Growth
Consider \(a_n=3n-2\). Evaluate it at several increasing indices:
The plotted points rise, and the table gives larger terms at larger indices. These observations suggest unbounded growth. The formula verifies that conclusion: for any fixed height \(H\), if \(n>\frac{H+2}{3}\), then \(3n-2>H\). Since indices can be chosen larger than this threshold, the sequence eventually exceeds \(H\), no matter how large the chosen height is. Thus the terms grow without bound rather than approaching a finite value. Substituting each listed index into \(3n-2\) reproduces the table values.
In an actual graph, a steep increase early on can make later growth look almost flat, or a compressed vertical scale can make large values look modest. Check the axis labels and compare numerical values as well as the apparent shape. The phrase “grows without bound” concerns what happens as the index continues to increase, not how tall the graph looks in the visible window.
Recognizing Persistent Oscillation
Some sequences move back and forth between separated values or bands. If this variation does not shrink toward one shared height, the plotted terms do not appear to settle to a single finite limit. It is important to distinguish that pattern from oscillation whose size shrinks: zigzagging by itself does not show that a sequence fails to approach a value.
Worked Example: Detecting Persistent Oscillation
Consider \(a_n=(-1)^n+\frac{1}{n}\). Compute several odd- and even-indexed terms:
The even-indexed terms are positive and near 1, while the odd-indexed terms after the first are negative and near \(-1\). The later entries remain in two separated bands, so the table suggests persistent oscillation rather than approach to one finite value. More directly, for any odd \(n\), the next term is at an even index, and their difference is
Since \(\frac{1}{n+1}<\frac{1}{n}\), this difference is less than 2, but it is still greater than 1: indeed, \(\frac{1}{n}-\frac{1}{n+1}=\frac{1}{n(n+1)}<1\), so the difference is greater than 1. Thus each such neighboring odd-even pair stays separated by more than 1, even at large indices. The plotted terms cannot all settle to one shared height. The substitutions above check the listed values, and the pair calculation explains why the separation persists.
Why Visual Evidence Is Useful—and Limited
Tables and graphs are valuable because they quickly reveal patterns worth investigating. A table lets you compare actual values at widely spaced indices. A graph can make a trend, a cluster, or a repeated swing apparent at a glance. Used together, they can help you decide what conclusion is plausible and which additional terms to inspect.
But finite evidence cannot settle every question. For instance, an initial run of increasing terms might be followed by a decrease. A handful of terms close to a particular number might later move away. A plotted window also hides all indices beyond its right edge. Treat the estimate as a conjecture to be checked, not as a proof based on the picture alone.
Check that the terms you compare really come from increasingly large values of \(n\).
Ask whether the terms cluster near one height, keep rising past new heights, or continue to occupy separated bands.
Use the formula, if available, to recompute suspicious entries and compare distances or changes between terms.
Say what the evidence suggests, and do not claim that a finite table or graph proves the behavior for all later indices.
A Common Mistake: Treating All Zigzags as Divergence
A sequence can alternate while still approaching a finite value if its swings shrink. For a concrete example, consider \(b_n=\frac{(-1)^n}{n}\). A mistaken conclusion would be, “The signs alternate, so the sequence does not have a limit.” The signs do alternate, but that fact alone does not determine the size of the terms.
The correction is to inspect the magnitudes: \(|b_n|=\frac{1}{n}\). These magnitudes decrease as \(n\) increases, so the plotted terms alternate above and below 0 while moving closer to 0. The evidence suggests a finite limit of 0, not persistent oscillation between separated bands. The calculation checks each displayed term and shows why the swings shrink. The mistake was to notice the alternating signs but ignore the changing size.
Use Estimates as a Starting Point
When examining a new sequence, combine several clues rather than relying on one visual feature. Compare values at increasing indices, inspect how far they are from a proposed finite value, and look for repeated patterns that do or do not shrink. If the formula is available, use it to test the estimate beyond the handful of entries in a table.
These habits lead to a useful preliminary classification: a likely finite limit, growth without bound, or persistent oscillation. They also prepare you to replace visual evidence with precise mathematical criteria. An estimate is a strong first step, but a proof is what establishes the behavior of every sufficiently late term.
Check Your Understanding
For each question, explain what the numerical or graphical evidence suggests, and distinguish an estimate from a proof.
- A table of terms at increasing indices stays closer to 4.2. What finite limit would you estimate, and what evidence would you inspect to check the estimate?
- The terms of \(c_n=5n+1\) are plotted. Describe how the formula confirms whether the sequence grows without bound.
- A graph alternates above and below the horizontal axis, but each point appears closer to 0. Why does alternation alone not establish persistent oscillation?
- What does a sequence graph plot, and why should connecting lines between its points not be treated as additional terms?
- Give one reason a finite table or graph cannot, by itself, prove the long-term behavior of a sequence.
